[ { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Campus area\", \"C\": \"Urban\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000000", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Campus area\", \"C\": \"Urban\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000000", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Dim streetlight visibility\", \"C\": \"Blinding sunlight\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000001", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Dim streetlight visibility\", \"C\": \"Blinding sunlight\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000001", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Sparse\", \"B\": \"Moderate\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000002", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Sparse\", \"B\": \"Moderate\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sparse" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000002", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Calm\", \"C\": \"Overcast\", \"D\": \"Dusty\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000003", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Calm\", \"C\": \"Overcast\", \"D\": \"Dusty\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Calm" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000003", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What color is the traffic light ahead? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Red\", \"C\": \"Yellow\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000004", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What color is the traffic light ahead? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Red\", \"C\": \"Yellow\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Red" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000004", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface currently? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Newly paved\", \"C\": \"Sandy\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000005", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/0.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface currently? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Newly paved\", \"C\": \"Sandy\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000005", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Rural\", \"C\": \"Construction zone\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000006", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Rural\", \"C\": \"Construction zone\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000006", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Low\", \"C\": \"Backlit\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000007", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Low\", \"C\": \"Backlit\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000007", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000008", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000008", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Smoggy\", \"C\": \"Dry\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000009", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Smoggy\", \"C\": \"Dry\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000009", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Countdown timer visible\", \"C\": \"None\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000010", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Countdown timer visible\", \"C\": \"None\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000010", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Snowy\", \"C\": \"Bumpy\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000011", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1006.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Snowy\", \"C\": \"Bumpy\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000011", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Shopping district\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000012", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Shopping district\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000012", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Strong shadows\", \"C\": \"Backlit\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000013", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Strong shadows\", \"C\": \"Backlit\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000013", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000014", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000014", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Windy\", \"C\": \"Strong wind\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000015", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Windy\", \"C\": \"Strong wind\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000015", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be observed? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Flashing yellow\", \"C\": \"Not visible\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000016", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be observed? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Flashing yellow\", \"C\": \"Not visible\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000016", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Partially damaged\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000017", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1015.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Partially damaged\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000017", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Urban\", \"C\": \"Airport roadway\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000018", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Urban\", \"C\": \"Airport roadway\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000018", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Hazy\", \"C\": \"Clear\", \"D\": \"Glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000019", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Hazy\", \"C\": \"Clear\", \"D\": \"Glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000019", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000020", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000020", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Thunderstorm\", \"C\": \"Clear\", \"D\": \"Drizzling\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000021", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Thunderstorm\", \"C\": \"Clear\", \"D\": \"Drizzling\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000021", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible and active? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Unknown\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000022", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible and active? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Unknown\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000022", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Dry\", \"C\": \"Leaf-covered\", \"D\": \"Uneven\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000023", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1024.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Dry\", \"C\": \"Leaf-covered\", \"D\": \"Uneven\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000023", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Campus area\", \"C\": \"Rural\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000024", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Campus area\", \"C\": \"Rural\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Highway" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000024", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Clear\", \"C\": \"Partly cloudy\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000025", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Clear\", \"C\": \"Partly cloudy\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000025", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Rush hour\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000026", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Rush hour\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000026", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Sandstorm\", \"C\": \"Windy\", \"D\": \"Drizzling\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000027", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Sandstorm\", \"C\": \"Windy\", \"D\": \"Drizzling\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000027", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Flashing red\", \"C\": \"Yellow\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000028", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Flashing red\", \"C\": \"Yellow\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000028", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Partially damaged\", \"C\": \"Smooth\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000029", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1033.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Partially damaged\", \"C\": \"Smooth\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000029", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Urban\", \"C\": \"Avenue\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000030", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Urban\", \"C\": \"Avenue\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Avenue" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000030", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dawn\", \"C\": \"Dim streetlight visibility\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000031", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dawn\", \"C\": \"Dim streetlight visibility\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000031", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000032", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000032", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Cloudy\", \"C\": \"Dusty\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000033", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Cloudy\", \"C\": \"Dusty\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000033", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Unknown\", \"C\": \"None\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000034", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Unknown\", \"C\": \"None\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000034", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Cracked asphalt\", \"C\": \"Leaf-covered\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000035", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1042.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Cracked asphalt\", \"C\": \"Leaf-covered\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000035", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Urban\", \"C\": \"Harbor area\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000036", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Urban\", \"C\": \"Harbor area\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000036", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Hazy\", \"C\": \"Poor\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000037", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Hazy\", \"C\": \"Poor\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000037", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000038", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000038", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Smoggy\", \"C\": \"Rainy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000039", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Smoggy\", \"C\": \"Rainy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000039", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Countdown timer visible\", \"C\": \"Flashing amber\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000040", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Countdown timer visible\", \"C\": \"Flashing amber\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000040", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Blocked by obstacle\", \"C\": \"Water puddles\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000041", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1051.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Blocked by obstacle\", \"C\": \"Water puddles\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000041", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Urban\", \"C\": \"Highway\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000042", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Urban\", \"C\": \"Highway\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000042", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Glare\", \"C\": \"Smoky\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000043", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Glare\", \"C\": \"Smoky\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000043", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000044", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000044", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Windy\", \"C\": \"Clear\", \"D\": \"Strong wind\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000045", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Windy\", \"C\": \"Clear\", \"D\": \"Strong wind\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000045", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Reflected signal\", \"C\": \"Green\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000046", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Reflected signal\", \"C\": \"Green\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000046", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000047", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1060.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000047", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Urban\", \"C\": \"Subway entrance\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000048", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Urban\", \"C\": \"Subway entrance\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000048", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Dawn\", \"C\": \"Low light\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000049", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Dawn\", \"C\": \"Low light\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000049", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000050", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000050", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Dusty\", \"C\": \"Windy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000051", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Dusty\", \"C\": \"Windy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000051", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state are the traffic lights likely in? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Green\", \"C\": \"Dim / unclear\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000052", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state are the traffic lights likely in? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Green\", \"C\": \"Dim / unclear\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000052", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Leaf-covered\", \"C\": \"Newly paved\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000053", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/107.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Leaf-covered\", \"C\": \"Newly paved\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000053", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Harbor area\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000054", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Harbor area\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000054", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Low\", \"C\": \"Overexposed\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000055", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Low\", \"C\": \"Overexposed\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000055", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000056", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000056", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Mist\", \"C\": \"Strong wind\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000057", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Mist\", \"C\": \"Strong wind\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000057", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the traffic lights ahead? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Green\", \"C\": \"Dim / unclear\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000058", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the traffic lights ahead? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Green\", \"C\": \"Dim / unclear\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000058", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Dry\", \"C\": \"Potholes\", \"D\": \"Icy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000059", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1079.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Dry\", \"C\": \"Potholes\", \"D\": \"Icy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000059", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Market street\", \"C\": \"Urban\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000060", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Market street\", \"C\": \"Urban\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000060", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Low\", \"C\": \"Flashing lights\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000061", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Low\", \"C\": \"Flashing lights\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000061", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Congested\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000062", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Congested\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000062", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Sandstorm\", \"C\": \"Dusty\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000063", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Sandstorm\", \"C\": \"Dusty\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000063", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state does the traffic signal appear to be in? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Covered by tree\", \"C\": \"Countdown timer visible\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000064", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state does the traffic signal appear to be in? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Covered by tree\", \"C\": \"Countdown timer visible\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000064", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Partially damaged\", \"C\": \"Potholes\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000065", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1088.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Partially damaged\", \"C\": \"Potholes\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000065", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Shopping district\", \"C\": \"Market street\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000066", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Shopping district\", \"C\": \"Market street\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Highway" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000066", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in the scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dim streetlight visibility\", \"C\": \"Dawn\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000067", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in the scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dim streetlight visibility\", \"C\": \"Dawn\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000067", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like in this scene? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000068", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like in this scene? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000068", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Foggy\", \"C\": \"Thunderstorm\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000069", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Foggy\", \"C\": \"Thunderstorm\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000069", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Countdown timer visible\", \"C\": \"Temporary road sign light\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000070", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Countdown timer visible\", \"C\": \"Temporary road sign light\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000070", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snowy\", \"C\": \"Bumpy\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000071", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1097.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snowy\", \"C\": \"Bumpy\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000071", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Urban\", \"C\": \"Forest road\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000072", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Urban\", \"C\": \"Forest road\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000072", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Overexposed\", \"C\": \"Flashing lights\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000073", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Overexposed\", \"C\": \"Flashing lights\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000073", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000074", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000074", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Sandstorm\", \"C\": \"Windy\", \"D\": \"Mist\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000075", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Sandstorm\", \"C\": \"Windy\", \"D\": \"Mist\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000075", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Pedestrian signal\", \"C\": \"Unknown\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000076", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Pedestrian signal\", \"C\": \"Unknown\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000076", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Bumpy\", \"C\": \"Rough surface\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000077", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1105.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Bumpy\", \"C\": \"Rough surface\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000077", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of street is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Market street\", \"C\": \"Tree-lined street\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000078", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of street is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Market street\", \"C\": \"Tree-lined street\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Tree-lined street" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000078", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Under tunnel\", \"C\": \"Backlit\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000079", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Under tunnel\", \"C\": \"Backlit\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000079", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000080", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000080", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Stormy\", \"C\": \"Smoggy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000081", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Stormy\", \"C\": \"Smoggy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000081", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible and active? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"None\", \"C\": \"Temporary road sign light\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000082", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible and active? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"None\", \"C\": \"Temporary road sign light\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000082", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Muddy\", \"C\": \"Dry\", \"D\": \"Icy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000083", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1114.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Muddy\", \"C\": \"Dry\", \"D\": \"Icy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000083", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Subway entrance\", \"C\": \"Airport roadway\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000084", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Subway entrance\", \"C\": \"Airport roadway\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000084", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low\", \"C\": \"Low light\", \"D\": \"Glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000085", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low\", \"C\": \"Low light\", \"D\": \"Glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000085", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000086", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000086", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Cloudy\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000087", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Cloudy\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000087", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Unknown\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000088", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Unknown\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000088", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000089", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1123.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000089", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Market street\", \"C\": \"Suburban\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000090", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Market street\", \"C\": \"Suburban\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000090", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Flashing lights\", \"C\": \"Rain-blurred\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000091", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Flashing lights\", \"C\": \"Rain-blurred\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000091", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000092", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000092", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Stormy\", \"C\": \"Clear\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000093", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Stormy\", \"C\": \"Clear\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000093", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Dim / unclear\", \"C\": \"Unknown\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000094", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Dim / unclear\", \"C\": \"Unknown\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000094", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Leaf-covered\", \"C\": \"Icy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000095", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1132.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Leaf-covered\", \"C\": \"Icy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000095", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Construction zone\", \"C\": \"Rural\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000096", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Construction zone\", \"C\": \"Rural\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000096", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Blinding sunlight\", \"C\": \"Backlit\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000097", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Blinding sunlight\", \"C\": \"Backlit\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000097", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in the scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000098", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in the scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000098", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is suggested in the scene? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Overcast\", \"C\": \"Flurries\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000099", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is suggested in the scene? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Overcast\", \"C\": \"Flurries\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000099", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined in this scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Flashing yellow\", \"C\": \"Flashing red\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000100", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined in this scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Flashing yellow\", \"C\": \"Flashing red\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000100", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear in this scene? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Oil slick\", \"C\": \"Snowy\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000101", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1141.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear in this scene? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Oil slick\", \"C\": \"Snowy\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000101", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Residential\", \"C\": \"Market street\", \"D\": \"Campus area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000102", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Residential\", \"C\": \"Market street\", \"D\": \"Campus area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000102", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Low\", \"C\": \"Reflected glare\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000103", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Low\", \"C\": \"Reflected glare\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000103", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are currently on the road? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"One\", \"C\": \"Moderate\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000104", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are currently on the road? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"One\", \"C\": \"Moderate\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000104", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Snowstorm\", \"C\": \"Hailing\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000105", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Snowstorm\", \"C\": \"Hailing\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000105", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Off / malfunctioning\", \"C\": \"Unknown\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000106", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Off / malfunctioning\", \"C\": \"Unknown\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000106", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Construction debris\", \"C\": \"Leaf-covered\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000107", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1150.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Construction debris\", \"C\": \"Leaf-covered\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000107", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Suburban\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000108", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Suburban\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000108", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Under tunnel\", \"C\": \"Low light\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000109", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Under tunnel\", \"C\": \"Low light\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000109", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000110", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000110", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Smoggy\", \"C\": \"Flurries\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000111", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Smoggy\", \"C\": \"Flurries\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000111", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What color is the traffic light visible in the distance? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Flashing red\", \"C\": \"Flashing amber\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000112", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What color is the traffic light visible in the distance? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Flashing red\", \"C\": \"Flashing amber\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Red" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000112", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Wet\", \"C\": \"Dry\", \"D\": \"Reflective\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000113", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/116.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Wet\", \"C\": \"Dry\", \"D\": \"Reflective\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Reflective" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000113", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Urban\", \"C\": \"Forest road\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000114", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Urban\", \"C\": \"Forest road\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000114", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Low\", \"C\": \"Partly cloudy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000115", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Low\", \"C\": \"Partly cloudy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000115", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000116", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000116", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Flurries\", \"C\": \"Strong wind\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000117", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Flurries\", \"C\": \"Strong wind\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000117", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the image? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Not visible\", \"C\": \"Red\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000118", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the image? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Not visible\", \"C\": \"Red\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000118", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Bumpy\", \"C\": \"Construction debris\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000119", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1169.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Bumpy\", \"C\": \"Construction debris\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000119", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Airport roadway\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000120", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Airport roadway\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000120", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Low\", \"C\": \"Reflected glare\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000121", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Low\", \"C\": \"Reflected glare\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000121", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Lane merging\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000122", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Lane merging\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000122", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Thunderstorm\", \"C\": \"Overcast\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000123", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Thunderstorm\", \"C\": \"Overcast\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000123", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the status of the traffic signal? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Covered by tree\", \"C\": \"Unknown\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000124", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the status of the traffic signal? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Covered by tree\", \"C\": \"Unknown\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000124", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Slippery\", \"C\": \"Dry\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000125", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1178.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Slippery\", \"C\": \"Dry\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000125", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Highway\", \"C\": \"Construction zone\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000126", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Highway\", \"C\": \"Construction zone\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000126", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Smoky\", \"C\": \"Poor\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000127", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Smoky\", \"C\": \"Poor\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000127", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000128", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000128", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Drizzling\", \"C\": \"Rainy\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000129", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Drizzling\", \"C\": \"Rainy\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000129", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Red\", \"C\": \"Reflected signal\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000130", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Red\", \"C\": \"Reflected signal\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000130", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Icy\", \"C\": \"Gravel\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000131", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1187.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Icy\", \"C\": \"Gravel\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000131", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Market street\", \"C\": \"Forest road\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000132", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Market street\", \"C\": \"Forest road\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000132", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Dawn\", \"C\": \"Hazy\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000133", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Dawn\", \"C\": \"Hazy\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000133", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000134", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000134", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Flurries\", \"C\": \"Windy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000135", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Flurries\", \"C\": \"Windy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000135", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Off / malfunctioning\", \"C\": \"Flashing red\", \"D\": \"Hidden\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000136", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Off / malfunctioning\", \"C\": \"Flashing red\", \"D\": \"Hidden\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000136", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Sandy\", \"C\": \"Dry\", \"D\": \"Potholes\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000137", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1196.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Sandy\", \"C\": \"Dry\", \"D\": \"Potholes\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000137", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Shopping district\", \"C\": \"Urban\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000138", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Shopping district\", \"C\": \"Urban\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000138", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Poor\", \"C\": \"Blinding sunlight\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000139", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Poor\", \"C\": \"Blinding sunlight\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000139", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000140", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000140", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Overcast\", \"C\": \"Clear\", \"D\": \"Drizzling\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000141", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Overcast\", \"C\": \"Clear\", \"D\": \"Drizzling\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000141", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Not visible\", \"C\": \"Pedestrian signal\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000142", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Not visible\", \"C\": \"Pedestrian signal\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000142", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000143", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1204.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000143", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Highway\", \"C\": \"Campus area\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000144", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Highway\", \"C\": \"Campus area\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000144", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Low\", \"C\": \"Blinding sunlight\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000145", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Low\", \"C\": \"Blinding sunlight\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000145", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000146", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000146", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Drizzling\", \"C\": \"Hailing\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000147", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Drizzling\", \"C\": \"Hailing\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000147", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing amber\", \"C\": \"Yellow\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000148", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing amber\", \"C\": \"Yellow\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000148", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Construction debris\", \"C\": \"Gravel\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000149", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1213.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Construction debris\", \"C\": \"Gravel\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000149", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000150", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000150", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Reflected glare\", \"C\": \"Poor\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000151", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Reflected glare\", \"C\": \"Poor\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000151", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000152", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000152", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Rainy\", \"C\": \"Strong wind\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000153", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Rainy\", \"C\": \"Strong wind\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000153", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"No\", \"C\": \"Reflected signal\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000154", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"No\", \"C\": \"Reflected signal\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000154", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Construction debris\", \"C\": \"Wet\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000155", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1222.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Construction debris\", \"C\": \"Wet\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000155", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Construction zone\", \"C\": \"Urban\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000156", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Construction zone\", \"C\": \"Urban\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000156", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Blinding sunlight\", \"C\": \"Poor\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000157", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Blinding sunlight\", \"C\": \"Poor\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000157", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Heavy\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000158", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Heavy\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000158", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Sandstorm\", \"C\": \"Rainy\", \"D\": \"Thunderstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000159", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Sandstorm\", \"C\": \"Rainy\", \"D\": \"Thunderstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000159", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"None\", \"C\": \"Green\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000160", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"None\", \"C\": \"Green\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000160", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Cracked asphalt\", \"C\": \"Slippery\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000161", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1231.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Cracked asphalt\", \"C\": \"Slippery\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000161", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Highway\", \"C\": \"Urban\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000162", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Highway\", \"C\": \"Urban\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000162", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Dawn\", \"C\": \"Glare\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000163", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Dawn\", \"C\": \"Glare\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000163", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000164", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000164", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Mist\", \"C\": \"Rain\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000165", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Mist\", \"C\": \"Rain\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rain" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000165", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing yellow\", \"C\": \"Yellow\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000166", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing yellow\", \"C\": \"Yellow\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000166", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Wet\", \"C\": \"Construction debris\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000167", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1240.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Wet\", \"C\": \"Construction debris\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000167", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000168", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000168", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Partly cloudy\", \"C\": \"Poor\", \"D\": \"Glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000169", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Partly cloudy\", \"C\": \"Poor\", \"D\": \"Glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000169", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000170", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000170", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Rainy\", \"C\": \"Smoggy\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000171", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Rainy\", \"C\": \"Smoggy\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000171", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Reflected signal\", \"C\": \"Unknown\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000172", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Reflected signal\", \"C\": \"Unknown\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000172", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Rough surface\", \"C\": \"Oil slick\", \"D\": \"Debris scattered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000173", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/125.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Rough surface\", \"C\": \"Oil slick\", \"D\": \"Debris scattered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000173", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Urban\", \"C\": \"Rural\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000174", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Urban\", \"C\": \"Rural\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000174", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Clear\", \"C\": \"Dawn\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000175", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Clear\", \"C\": \"Dawn\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000175", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Rush hour\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000176", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Rush hour\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000176", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Strong wind\", \"C\": \"Flurries\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000177", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Strong wind\", \"C\": \"Flurries\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000177", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Dim / unclear\", \"C\": \"Unknown\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000178", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Dim / unclear\", \"C\": \"Unknown\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000178", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Wet\", \"C\": \"Cracked asphalt\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000179", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1259.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Wet\", \"C\": \"Cracked asphalt\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000179", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Urban\", \"C\": \"Market street\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000180", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Urban\", \"C\": \"Market street\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000180", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Partly cloudy\", \"C\": \"Dusk\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000181", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Partly cloudy\", \"C\": \"Dusk\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000181", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000182", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000182", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Overcast\", \"C\": \"Snowy\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000183", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Overcast\", \"C\": \"Snowy\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000183", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"No\", \"C\": \"Flashing red\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000184", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"No\", \"C\": \"Flashing red\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000184", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Wet\", \"C\": \"Bumpy\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000185", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1268.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Wet\", \"C\": \"Bumpy\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000185", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Suburban\", \"C\": \"Construction zone\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000186", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Suburban\", \"C\": \"Construction zone\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000186", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Poor\", \"C\": \"Smoky\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000187", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Poor\", \"C\": \"Smoky\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000187", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000188", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000188", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Dusty\", \"C\": \"Sleet\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000189", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Dusty\", \"C\": \"Sleet\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000189", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Off / malfunctioning\", \"C\": \"Green\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000190", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Off / malfunctioning\", \"C\": \"Green\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000190", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Smooth\", \"C\": \"Water puddles\", \"D\": \"Icy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000191", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1277.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Smooth\", \"C\": \"Water puddles\", \"D\": \"Icy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000191", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this drive occurring in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Rural\", \"C\": \"Forest road\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000192", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this drive occurring in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Rural\", \"C\": \"Forest road\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000192", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Flashing lights\", \"C\": \"Blinding sunlight\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000193", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Flashing lights\", \"C\": \"Blinding sunlight\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000193", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000194", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000194", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Mist\", \"D\": \"Overcast\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000195", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Mist\", \"D\": \"Overcast\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000195", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Covered by tree\", \"C\": \"Flashing red\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000196", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Covered by tree\", \"C\": \"Flashing red\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000196", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000197", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1286.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000197", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Airport roadway\", \"C\": \"Campus area\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000198", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Airport roadway\", \"C\": \"Campus area\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000198", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Low\", \"C\": \"Blinding sunlight\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000199", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Low\", \"C\": \"Blinding sunlight\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000199", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the scene? You should answer one choice from four answers \"A\": \"One\", \"B\": \"Lane merging\", \"C\": \"Rush hour\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000200", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the scene? You should answer one choice from four answers \"A\": \"One\", \"B\": \"Lane merging\", \"C\": \"Rush hour\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000200", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Clear\", \"C\": \"Sandstorm\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000201", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Clear\", \"C\": \"Sandstorm\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000201", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000202", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000202", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Leaf-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000203", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1295.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Leaf-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000203", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Campus area\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000204", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Campus area\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000204", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Poor\", \"C\": \"Strong shadows\", \"D\": \"Under tunnel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000205", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Poor\", \"C\": \"Strong shadows\", \"D\": \"Under tunnel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000205", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the traffic? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"One\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000206", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the traffic? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"One\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000206", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Clear\", \"C\": \"Snowstorm\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000207", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Clear\", \"C\": \"Snowstorm\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000207", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Flashing yellow\", \"C\": \"Dim / unclear\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000208", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Flashing yellow\", \"C\": \"Dim / unclear\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000208", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Partially damaged\", \"C\": \"Newly paved\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000209", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1303.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Partially damaged\", \"C\": \"Newly paved\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000209", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Parking lot\", \"C\": \"Campus area\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000210", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Parking lot\", \"C\": \"Campus area\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000210", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Overexposed\", \"C\": \"Dawn\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000211", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Overexposed\", \"C\": \"Dawn\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000211", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000212", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000212", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000213", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000213", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Yellow\", \"C\": \"Temporary road sign light\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000214", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Yellow\", \"C\": \"Temporary road sign light\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000214", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000215", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1312.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000215", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Parking lot\", \"C\": \"Construction zone\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000216", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Parking lot\", \"C\": \"Construction zone\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000216", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Rain-blurred\", \"C\": \"Low\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000217", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Rain-blurred\", \"C\": \"Low\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000217", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000218", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000218", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Dry\", \"C\": \"Stormy\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000219", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Dry\", \"C\": \"Stormy\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000219", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Hidden\", \"B\": \"Temporary road sign light\", \"C\": \"Blinking green\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000220", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Hidden\", \"B\": \"Temporary road sign light\", \"C\": \"Blinking green\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000220", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Dry\", \"C\": \"Oil slick\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000221", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1321.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Dry\", \"C\": \"Oil slick\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000221", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Construction zone\", \"C\": \"Suburban\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000222", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Construction zone\", \"C\": \"Suburban\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000222", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Poor\", \"C\": \"Reflected glare\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000223", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Poor\", \"C\": \"Reflected glare\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000223", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000224", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000224", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Overcast\", \"C\": \"Rainy\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000225", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Overcast\", \"C\": \"Rainy\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000225", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Reflected signal\", \"C\": \"Yellow\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000226", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Reflected signal\", \"C\": \"Yellow\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000226", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Uneven\", \"C\": \"Debris scattered\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000227", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1330.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Uneven\", \"C\": \"Debris scattered\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000227", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Construction zone\", \"C\": \"Urban\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000228", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Construction zone\", \"C\": \"Urban\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000228", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Low\", \"C\": \"Foggy\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000229", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Low\", \"C\": \"Foggy\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000229", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Heavy\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000230", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Heavy\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000230", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Mist\", \"C\": \"Windy\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000231", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Mist\", \"C\": \"Windy\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000231", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the likely state of the traffic signal? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Dim / unclear\", \"C\": \"Temporary road sign light\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000232", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the likely state of the traffic signal? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Dim / unclear\", \"C\": \"Temporary road sign light\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000232", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Smooth\", \"C\": \"Debris scattered\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000233", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/134.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Smooth\", \"C\": \"Debris scattered\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000233", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Roundabout\", \"C\": \"Coastal road\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000234", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Roundabout\", \"C\": \"Coastal road\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000234", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Dusk\", \"C\": \"Under tunnel\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000235", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Dusk\", \"C\": \"Under tunnel\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000235", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000236", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000236", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Gloomy\", \"C\": \"Dusty\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000237", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Gloomy\", \"C\": \"Dusty\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000237", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"None\", \"C\": \"Red\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000238", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"None\", \"C\": \"Red\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000238", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Cracked asphalt\", \"C\": \"Newly paved\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000239", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1349.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Cracked asphalt\", \"C\": \"Newly paved\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000239", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Construction zone\", \"C\": \"Parkway\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000240", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Construction zone\", \"C\": \"Parkway\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Parkway" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000240", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Clear\", \"C\": \"Flashing lights\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000241", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Clear\", \"C\": \"Flashing lights\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000241", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000242", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000242", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Flurries\", \"C\": \"Gloomy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000243", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Flurries\", \"C\": \"Gloomy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000243", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Unknown\", \"C\": \"Reflected signal\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000244", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Unknown\", \"C\": \"Reflected signal\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000244", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Dry\", \"C\": \"Rough surface\", \"D\": \"Potholes\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000245", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1358.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Dry\", \"C\": \"Rough surface\", \"D\": \"Potholes\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000245", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Urban\", \"C\": \"Rural\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000246", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Urban\", \"C\": \"Rural\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000246", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Dawn\", \"C\": \"Glare\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000247", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Dawn\", \"C\": \"Glare\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000247", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Heavy\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000248", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Heavy\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000248", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Rainy\", \"C\": \"Rain\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000249", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Rainy\", \"C\": \"Rain\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rain" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000249", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Green\", \"C\": \"Yellow\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000250", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Green\", \"C\": \"Yellow\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000250", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Leaf-covered\", \"C\": \"Water puddles\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000251", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1367.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Leaf-covered\", \"C\": \"Water puddles\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000251", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Bridge\", \"C\": \"Roundabout\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000252", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Bridge\", \"C\": \"Roundabout\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000252", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Blinding sunlight\", \"C\": \"Foggy\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000253", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Blinding sunlight\", \"C\": \"Foggy\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000253", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Lane merging\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000254", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Lane merging\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000254", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Dusty\", \"C\": \"Rainy\", \"D\": \"Overcast\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000255", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Dusty\", \"C\": \"Rainy\", \"D\": \"Overcast\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000255", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Flashing amber\", \"C\": \"Yellow\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000256", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Flashing amber\", \"C\": \"Yellow\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000256", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Oil slick\", \"C\": \"Construction debris\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000257", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1376.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Oil slick\", \"C\": \"Construction debris\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000257", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Market street\", \"C\": \"Suburban\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000258", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Market street\", \"C\": \"Suburban\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000258", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Smoky\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000259", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Smoky\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000259", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic situation like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000260", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic situation like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000260", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Mist\", \"C\": \"Rainy\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000261", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Mist\", \"C\": \"Rainy\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000261", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Pedestrian signal\", \"C\": \"Temporary road sign light\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000262", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Pedestrian signal\", \"C\": \"Temporary road sign light\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000262", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Wet\", \"C\": \"Snowy\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000263", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1385.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Wet\", \"C\": \"Snowy\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000263", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Suburban\", \"C\": \"Service station\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000264", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Suburban\", \"C\": \"Service station\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000264", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Overexposed\", \"C\": \"Hazy\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000265", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Overexposed\", \"C\": \"Hazy\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000265", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity observed? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000266", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity observed? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000266", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is currently present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Hailing\", \"C\": \"Rainy\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000267", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is currently present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Hailing\", \"C\": \"Rainy\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000267", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the scene? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Red\", \"C\": \"Flashing yellow\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000268", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the scene? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Red\", \"C\": \"Flashing yellow\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000268", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Potholes\", \"C\": \"Oil slick\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000269", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1394.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Potholes\", \"C\": \"Oil slick\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000269", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000270", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000270", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Limited\", \"B\": \"Clear\", \"C\": \"Dim streetlight visibility\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000271", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Limited\", \"B\": \"Clear\", \"C\": \"Dim streetlight visibility\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Limited" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000271", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000272", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000272", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Smoggy\", \"C\": \"Clear\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000273", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Smoggy\", \"C\": \"Clear\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000273", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined in this scene? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing yellow\", \"C\": \"Flashing red\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000274", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined in this scene? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing yellow\", \"C\": \"Flashing red\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000274", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described in the scene? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Rough surface\", \"C\": \"Dry\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000275", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/1402.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described in the scene? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Rough surface\", \"C\": \"Dry\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000275", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Bridge\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000276", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Bridge\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000276", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Clear\", \"C\": \"Dawn\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000277", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Clear\", \"C\": \"Dawn\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000277", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000278", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000278", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Rainy\", \"C\": \"Dusty\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000279", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Rainy\", \"C\": \"Dusty\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000279", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state are the traffic lights likely in? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Temporary road sign light\", \"C\": \"Yellow\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000280", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state are the traffic lights likely in? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Temporary road sign light\", \"C\": \"Yellow\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000280", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Wet\", \"C\": \"Partially damaged\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000281", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/148.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Wet\", \"C\": \"Partially damaged\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000281", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Highway\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000282", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Highway\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000282", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Poor\", \"C\": \"Under tunnel\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000283", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Poor\", \"C\": \"Under tunnel\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000283", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Heavy\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000284", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Heavy\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000284", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Rainy\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000285", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Rainy\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000285", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What can be inferred about the traffic signal ahead? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Pedestrian signal\", \"C\": \"Unknown\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000286", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What can be inferred about the traffic signal ahead? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Pedestrian signal\", \"C\": \"Unknown\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000286", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Cracked asphalt\", \"C\": \"Partially damaged\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000287", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/157.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Cracked asphalt\", \"C\": \"Partially damaged\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000287", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Roundabout\", \"C\": \"Rural\", \"D\": \"Curved rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000288", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Roundabout\", \"C\": \"Rural\", \"D\": \"Curved rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Curved rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000288", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Partly cloudy\", \"C\": \"Dim streetlight visibility\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000289", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Partly cloudy\", \"C\": \"Dim streetlight visibility\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000289", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000290", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000290", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Thunderstorm\", \"C\": \"Hailing\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000291", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Thunderstorm\", \"C\": \"Hailing\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000291", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing red\", \"C\": \"Red\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000292", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing red\", \"C\": \"Red\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000292", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface condition? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Icy\", \"C\": \"Newly paved\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000293", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/166.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface condition? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Icy\", \"C\": \"Newly paved\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000293", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000294", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000294", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Rain-blurred\", \"C\": \"Foggy\", \"D\": \"Blinding sunlight\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000295", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Rain-blurred\", \"C\": \"Foggy\", \"D\": \"Blinding sunlight\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000295", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000296", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000296", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Hailing\", \"C\": \"Rainy\", \"D\": \"Windy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000297", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Hailing\", \"C\": \"Rainy\", \"D\": \"Windy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000297", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Flashing yellow\", \"C\": \"Countdown timer visible\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000298", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Flashing yellow\", \"C\": \"Countdown timer visible\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000298", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Flooded\", \"C\": \"Wet\", \"D\": \"Cracked asphalt\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000299", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/175.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Flooded\", \"C\": \"Wet\", \"D\": \"Cracked asphalt\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000299", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Curved road\", \"C\": \"Campus area\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000300", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Curved road\", \"C\": \"Campus area\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Curved road" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000300", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Overexposed\", \"C\": \"Partly cloudy\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000301", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Overexposed\", \"C\": \"Partly cloudy\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000301", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible ahead? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"One\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000302", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible ahead? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"One\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000302", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Flurries\", \"C\": \"Overcast\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000303", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Flurries\", \"C\": \"Overcast\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Overcast" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000303", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic signal visible? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Not visible\", \"C\": \"Temporary road sign light\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000304", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic signal visible? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Not visible\", \"C\": \"Temporary road sign light\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000304", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Dry\", \"C\": \"Rough surface\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000305", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/184.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Dry\", \"C\": \"Rough surface\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000305", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Airport roadway\", \"C\": \"Subway entrance\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000306", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Airport roadway\", \"C\": \"Subway entrance\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000306", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Smoky\", \"C\": \"Under tunnel\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000307", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Smoky\", \"C\": \"Under tunnel\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000307", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic situation in the lane? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000308", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic situation in the lane? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000308", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Drizzling\", \"C\": \"Sleet\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000309", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Drizzling\", \"C\": \"Sleet\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000309", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Flashing amber\", \"C\": \"Unknown\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000310", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Flashing amber\", \"C\": \"Unknown\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000310", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Flooded\", \"C\": \"Smooth\", \"D\": \"Uneven\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000311", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/193.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Flooded\", \"C\": \"Smooth\", \"D\": \"Uneven\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000311", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Parking lot\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000312", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Parking lot\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000312", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Overexposed\", \"C\": \"Clear\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000313", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Overexposed\", \"C\": \"Clear\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000313", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are moving? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000314", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are moving? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000314", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather like? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Flurries\", \"C\": \"Mist\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000315", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather like? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Flurries\", \"C\": \"Mist\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000315", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic signal state? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Yellow\", \"C\": \"Temporary road sign light\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000316", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic signal state? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Yellow\", \"C\": \"Temporary road sign light\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000316", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Muddy\", \"C\": \"Sandy\", \"D\": \"Leaf-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000317", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/201.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Muddy\", \"C\": \"Sandy\", \"D\": \"Leaf-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000317", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Bridge\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000318", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Bridge\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000318", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Hazy\", \"C\": \"Under tunnel\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000319", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Hazy\", \"C\": \"Under tunnel\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000319", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000320", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000320", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Strong wind\", \"C\": \"Gloomy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000321", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Strong wind\", \"C\": \"Gloomy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000321", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Not visible\", \"C\": \"Flashing yellow\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000322", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Not visible\", \"C\": \"Flashing yellow\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000322", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Muddy\", \"C\": \"Wet\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000323", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/210.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Muddy\", \"C\": \"Wet\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000323", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Forest road\", \"C\": \"Subway entrance\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000324", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Forest road\", \"C\": \"Subway entrance\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000324", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Dusk\", \"C\": \"Low\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000325", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Dusk\", \"C\": \"Low\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000325", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000326", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000326", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Smoggy\", \"C\": \"Rainy\", \"D\": \"Thunderstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000327", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Smoggy\", \"C\": \"Rainy\", \"D\": \"Thunderstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000327", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Unknown\", \"C\": \"Off / malfunctioning\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000328", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Unknown\", \"C\": \"Off / malfunctioning\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000328", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Partially damaged\", \"C\": \"Dry\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000329", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/22.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Partially damaged\", \"C\": \"Dry\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000329", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000330", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000330", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Foggy\", \"C\": \"Glare\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000331", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Foggy\", \"C\": \"Glare\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000331", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000332", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000332", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Strong wind\", \"C\": \"Snowstorm\", \"D\": \"Windy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000333", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Strong wind\", \"C\": \"Snowstorm\", \"D\": \"Windy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000333", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state are the traffic signals likely in? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Flashing amber\", \"C\": \"Pedestrian signal\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000334", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state are the traffic signals likely in? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Flashing amber\", \"C\": \"Pedestrian signal\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000334", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Uneven\", \"C\": \"Debris scattered\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000335", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/229.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Uneven\", \"C\": \"Debris scattered\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000335", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Bridge\", \"C\": \"Coastal road\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000336", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Bridge\", \"C\": \"Coastal road\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Highway" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000336", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Rain-blurred\", \"C\": \"Strong shadows\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000337", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Rain-blurred\", \"C\": \"Strong shadows\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000337", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Moderate\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000338", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Moderate\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000338", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Stormy\", \"C\": \"Clear\", \"D\": \"Dusty\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000339", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Stormy\", \"C\": \"Clear\", \"D\": \"Dusty\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000339", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Temporary road sign light\", \"C\": \"Unknown\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000340", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Temporary road sign light\", \"C\": \"Unknown\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000340", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Potholes\", \"C\": \"Water puddles\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000341", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/238.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Potholes\", \"C\": \"Water puddles\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000341", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Urban\", \"C\": \"Subway entrance\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000342", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Urban\", \"C\": \"Subway entrance\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000342", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Low light\", \"C\": \"Dawn\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000343", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Low light\", \"C\": \"Dawn\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000343", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000344", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000344", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Dry\", \"C\": \"Windy\", \"D\": \"Thunderstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000345", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Dry\", \"C\": \"Windy\", \"D\": \"Thunderstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000345", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Yellow\", \"C\": \"Temporary road sign light\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000346", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Yellow\", \"C\": \"Temporary road sign light\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000346", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Debris scattered\", \"C\": \"Rough surface\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000347", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/247.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Debris scattered\", \"C\": \"Rough surface\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000347", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Urban\", \"C\": \"Suburban\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000348", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Urban\", \"C\": \"Suburban\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000348", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Dawn\", \"C\": \"Dusk\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000349", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Dawn\", \"C\": \"Dusk\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000349", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000350", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000350", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Cloudy\", \"C\": \"Clear\", \"D\": \"Mist\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000351", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Cloudy\", \"C\": \"Clear\", \"D\": \"Mist\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000351", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Flashing amber\", \"C\": \"Unknown\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000352", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Flashing amber\", \"C\": \"Unknown\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000352", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Icy\", \"C\": \"Leaf-covered\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000353", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/256.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Icy\", \"C\": \"Leaf-covered\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000353", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this drive in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Forest road\", \"C\": \"Service station\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000354", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this drive in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Forest road\", \"C\": \"Service station\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000354", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Blinding sunlight\", \"C\": \"Overexposed\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000355", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Blinding sunlight\", \"C\": \"Overexposed\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000355", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000356", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000356", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dusty\", \"C\": \"Cloudy\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000357", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dusty\", \"C\": \"Cloudy\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000357", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Not visible\", \"C\": \"Dim / unclear\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000358", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Not visible\", \"C\": \"Dim / unclear\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000358", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface in this scene? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Uneven\", \"C\": \"Dry\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000359", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/265.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface in this scene? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Uneven\", \"C\": \"Dry\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000359", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Harbor area\", \"C\": \"Roundabout\", \"D\": \"Campus area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000360", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Harbor area\", \"C\": \"Roundabout\", \"D\": \"Campus area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000360", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Hazy\", \"C\": \"Dim streetlight visibility\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000361", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Hazy\", \"C\": \"Dim streetlight visibility\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000361", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000362", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000362", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Sandstorm\", \"C\": \"Stormy\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000363", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Sandstorm\", \"C\": \"Stormy\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000363", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000364", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000364", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current surface condition? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Bumpy\", \"C\": \"Rough surface\", \"D\": \"Gravel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000365", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/274.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current surface condition? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Bumpy\", \"C\": \"Rough surface\", \"D\": \"Gravel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000365", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Bridge\", \"C\": \"Shopping district\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000366", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Bridge\", \"C\": \"Shopping district\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Highway" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000366", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Glare\", \"C\": \"Reflected glare\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000367", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Glare\", \"C\": \"Reflected glare\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000367", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000368", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000368", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Strong wind\", \"C\": \"Sunny\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000369", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Strong wind\", \"C\": \"Sunny\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000369", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Not visible\", \"C\": \"Green\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000370", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Not visible\", \"C\": \"Green\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000370", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Uneven\", \"C\": \"Smooth\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000371", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/283.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Uneven\", \"C\": \"Smooth\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000371", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Highway\", \"C\": \"Construction zone\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000372", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Highway\", \"C\": \"Construction zone\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000372", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Clear\", \"C\": \"Backlit\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000373", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Clear\", \"C\": \"Backlit\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000373", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition ahead? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000374", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition ahead? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000374", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Flurries\", \"C\": \"Sleet\", \"D\": \"Windy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000375", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Flurries\", \"C\": \"Sleet\", \"D\": \"Windy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000375", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Off / malfunctioning\", \"C\": \"Reflected signal\", \"D\": \"Hidden\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000376", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Off / malfunctioning\", \"C\": \"Reflected signal\", \"D\": \"Hidden\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000376", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Icy\", \"C\": \"Sandy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000377", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/292.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Icy\", \"C\": \"Sandy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000377", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this drive occurring in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Market street\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000378", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this drive occurring in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Market street\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000378", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Smoky\", \"C\": \"Partly cloudy\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000379", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Smoky\", \"C\": \"Partly cloudy\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000379", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in motion? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"One\", \"C\": \"Moderate\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000380", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in motion? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"One\", \"C\": \"Moderate\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000380", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Cloudy\", \"C\": \"Stormy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000381", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Cloudy\", \"C\": \"Stormy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000381", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"None\", \"C\": \"Not visible\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000382", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"None\", \"C\": \"Not visible\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000382", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Dry\", \"C\": \"Gravel\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000383", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/300.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Dry\", \"C\": \"Gravel\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000383", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Highway\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000384", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Highway\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000384", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Foggy\", \"C\": \"Strong shadows\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000385", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Foggy\", \"C\": \"Strong shadows\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000385", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Accident blockage\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000386", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Accident blockage\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000386", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Snowstorm\", \"C\": \"Clear\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000387", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Snowstorm\", \"C\": \"Clear\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000387", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the image? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Flashing yellow\", \"C\": \"Temporary road sign light\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000388", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the image? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Flashing yellow\", \"C\": \"Temporary road sign light\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000388", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Debris scattered\", \"C\": \"Partially damaged\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000389", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/31.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Debris scattered\", \"C\": \"Partially damaged\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000389", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Semi-urban\", \"B\": \"Market street\", \"C\": \"Highway\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000390", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Semi-urban\", \"B\": \"Market street\", \"C\": \"Highway\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Semi-urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000390", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Partly cloudy\", \"C\": \"Low\", \"D\": \"Blinding sunlight\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000391", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Partly cloudy\", \"C\": \"Low\", \"D\": \"Blinding sunlight\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000391", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000392", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000392", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Clear\", \"C\": \"Foggy\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000393", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Clear\", \"C\": \"Foggy\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000393", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Covered by tree\", \"C\": \"Red\", \"D\": \"No\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000394", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Covered by tree\", \"C\": \"Red\", \"D\": \"No\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000394", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Oil slick\", \"C\": \"Smooth\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000395", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/319.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Oil slick\", \"C\": \"Smooth\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000395", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Campus area\", \"C\": \"Market street\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000396", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Campus area\", \"C\": \"Market street\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000396", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road visibility affected at night? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Rain-blurred\", \"C\": \"Dawn\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000397", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road visibility affected at night? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Rain-blurred\", \"C\": \"Dawn\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000397", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"One\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000398", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"One\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000398", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Dusty\", \"C\": \"Clear\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000399", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Dusty\", \"C\": \"Clear\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000399", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic signal visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Flashing red\", \"C\": \"No\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000400", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic signal visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Flashing red\", \"C\": \"No\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000400", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Gravel\", \"C\": \"Dry\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000401", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/328.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Gravel\", \"C\": \"Dry\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000401", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this drive taking place in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Bridge\", \"C\": \"Suburban\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000402", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this drive taking place in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Bridge\", \"C\": \"Suburban\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000402", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Strong shadows\", \"C\": \"Poor\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000403", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Strong shadows\", \"C\": \"Poor\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000403", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000404", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000404", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Clear\", \"C\": \"Mist\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000405", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Clear\", \"C\": \"Mist\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000405", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Countdown timer visible\", \"C\": \"Not visible\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000406", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Countdown timer visible\", \"C\": \"Not visible\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000406", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rough surface\", \"C\": \"Blocked by obstacle\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000407", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/337.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rough surface\", \"C\": \"Blocked by obstacle\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000407", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Service station\", \"C\": \"Subway entrance\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000408", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Service station\", \"C\": \"Subway entrance\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000408", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Foggy\", \"C\": \"Dusk\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000409", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Foggy\", \"C\": \"Dusk\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000409", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000410", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000410", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hailing\", \"C\": \"Drizzling\", \"D\": \"Mist\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000411", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hailing\", \"C\": \"Drizzling\", \"D\": \"Mist\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000411", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this frame? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Green\", \"C\": \"Yellow\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000412", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this frame? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Green\", \"C\": \"Yellow\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000412", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000413", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/346.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000413", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Campus area\", \"C\": \"Service station\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000414", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Campus area\", \"C\": \"Service station\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000414", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the image quality in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Smoky\", \"C\": \"Reflected glare\", \"D\": \"Blurry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000415", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the image quality in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Smoky\", \"C\": \"Reflected glare\", \"D\": \"Blurry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Blurry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000415", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000416", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000416", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Drizzling\", \"C\": \"Stormy\", \"D\": \"Sleet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000417", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Drizzling\", \"C\": \"Stormy\", \"D\": \"Sleet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000417", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state do the traffic signals appear to be in? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Green\", \"C\": \"Pedestrian signal\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000418", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state do the traffic signals appear to be in? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Green\", \"C\": \"Pedestrian signal\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000418", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Debris scattered\", \"C\": \"Dry\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000419", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/355.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Debris scattered\", \"C\": \"Dry\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000419", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Tree-lined street\", \"B\": \"Suburban\", \"C\": \"Campus area\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000420", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Tree-lined street\", \"B\": \"Suburban\", \"C\": \"Campus area\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Tree-lined street" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000420", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene described? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Clear\", \"C\": \"Rain-blurred\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000421", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene described? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Clear\", \"C\": \"Rain-blurred\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000421", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like in this scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000422", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like in this scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000422", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Wet\", \"C\": \"Mist\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000423", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Wet\", \"C\": \"Mist\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000423", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Not visible\", \"C\": \"Yellow\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000424", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Not visible\", \"C\": \"Yellow\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000424", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Smooth\", \"C\": \"Cracked asphalt\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000425", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/364.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Smooth\", \"C\": \"Cracked asphalt\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000425", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000426", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000426", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Hazy\", \"C\": \"Low\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000427", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Hazy\", \"C\": \"Low\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000427", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000428", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000428", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Drizzling\", \"C\": \"Windy\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000429", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Drizzling\", \"C\": \"Windy\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000429", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Reflected signal\", \"C\": \"Covered by tree\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000430", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Reflected signal\", \"C\": \"Covered by tree\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000430", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Muddy\", \"C\": \"Potholes\", \"D\": \"Reflective\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000431", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/373.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Muddy\", \"C\": \"Potholes\", \"D\": \"Reflective\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Reflective" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000431", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Shopping district\", \"C\": \"Highway\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000432", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Shopping district\", \"C\": \"Highway\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000432", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Under tunnel\", \"C\": \"Rain-blurred\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000433", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Under tunnel\", \"C\": \"Rain-blurred\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000433", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Moderate\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000434", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Moderate\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000434", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Rainy\", \"C\": \"Hailing\", \"D\": \"Overcast\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000435", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Rainy\", \"C\": \"Hailing\", \"D\": \"Overcast\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000435", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Temporary road sign light\", \"C\": \"Not visible\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000436", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Temporary road sign light\", \"C\": \"Not visible\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000436", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Wet\", \"C\": \"Partially damaged\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000437", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/382.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Wet\", \"C\": \"Partially damaged\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000437", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Highway\", \"C\": \"Subway entrance\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000438", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Highway\", \"C\": \"Subway entrance\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000438", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Strong shadows\", \"C\": \"Low light\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000439", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Strong shadows\", \"C\": \"Low light\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000439", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Accident blockage\", \"C\": \"Heavy\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000440", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Accident blockage\", \"C\": \"Heavy\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000440", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Sunny\", \"C\": \"Cloudy\", \"D\": \"Overcast\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000441", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Sunny\", \"C\": \"Cloudy\", \"D\": \"Overcast\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000441", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Hidden\", \"B\": \"Temporary road sign light\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000442", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Hidden\", \"B\": \"Temporary road sign light\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000442", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Wet\", \"C\": \"Blocked by obstacle\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000443", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/391.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Wet\", \"C\": \"Blocked by obstacle\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000443", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Shopping district\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000444", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Shopping district\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000444", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Clear\", \"C\": \"Dusk\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000445", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Clear\", \"C\": \"Dusk\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000445", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000446", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000446", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Clear\", \"C\": \"Overcast\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000447", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Clear\", \"C\": \"Overcast\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000447", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state do the traffic signals appear to be in? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Off / malfunctioning\", \"C\": \"Red\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000448", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state do the traffic signals appear to be in? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Off / malfunctioning\", \"C\": \"Red\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000448", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Rough surface\", \"C\": \"Smooth\", \"D\": \"Leaf-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000449", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/40.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Rough surface\", \"C\": \"Smooth\", \"D\": \"Leaf-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000449", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of street is depicted in the scene? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Parking lot\", \"C\": \"Airport roadway\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000450", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of street is depicted in the scene? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Parking lot\", \"C\": \"Airport roadway\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000450", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Rain-blurred\", \"C\": \"Low\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000451", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Rain-blurred\", \"C\": \"Low\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000451", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000452", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000452", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Smoggy\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000453", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Smoggy\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000453", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Countdown timer visible\", \"C\": \"None\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000454", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Countdown timer visible\", \"C\": \"None\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000454", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Dry\", \"C\": \"Water puddles\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000455", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/409.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Dry\", \"C\": \"Water puddles\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000455", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Subway entrance\", \"C\": \"Parking lot\", \"D\": \"Two-lane road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000456", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Subway entrance\", \"C\": \"Parking lot\", \"D\": \"Two-lane road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Two-lane road" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000456", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility affected in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Glare\", \"C\": \"Poor\", \"D\": \"Reflected glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000457", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility affected in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Glare\", \"C\": \"Poor\", \"D\": \"Reflected glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000457", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Accident blockage\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000458", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Accident blockage\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000458", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Flurries\", \"C\": \"Rainy\", \"D\": \"Thunderstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000459", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Flurries\", \"C\": \"Rainy\", \"D\": \"Thunderstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000459", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Off / malfunctioning\", \"C\": \"Flashing yellow\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000460", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Off / malfunctioning\", \"C\": \"Flashing yellow\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000460", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Snowy\", \"C\": \"Snow-covered\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000461", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/418.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Snowy\", \"C\": \"Snow-covered\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000461", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Avenue\", \"C\": \"Parking lot\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000462", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Avenue\", \"C\": \"Parking lot\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Avenue" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000462", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility affected in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Flashing lights\", \"C\": \"Low\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000463", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility affected in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Flashing lights\", \"C\": \"Low\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000463", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000464", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000464", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Drizzling\", \"C\": \"Rainy\", \"D\": \"Windy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000465", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Drizzling\", \"C\": \"Rainy\", \"D\": \"Windy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000465", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Unknown\", \"C\": \"Reflected signal\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000466", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Unknown\", \"C\": \"Reflected signal\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000466", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Muddy\", \"C\": \"Damp\", \"D\": \"Leaf-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000467", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/427.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Muddy\", \"C\": \"Damp\", \"D\": \"Leaf-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Damp" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000467", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Forest road\", \"C\": \"Coastal road\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000468", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Forest road\", \"C\": \"Coastal road\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Highway" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000468", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low light\", \"C\": \"Strong shadows\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000469", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low light\", \"C\": \"Strong shadows\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000469", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Congested\", \"C\": \"Moderate\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000470", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Congested\", \"C\": \"Moderate\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000470", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Clear\", \"C\": \"Flurries\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000471", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Clear\", \"C\": \"Flurries\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000471", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Not visible\", \"C\": \"Blinking green\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000472", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Not visible\", \"C\": \"Blinking green\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000472", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Wet\", \"C\": \"Potholes\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000473", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/436.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Wet\", \"C\": \"Potholes\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000473", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Service station\", \"C\": \"Construction zone\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000474", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Service station\", \"C\": \"Construction zone\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000474", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Overexposed\", \"C\": \"Poor\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000475", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Overexposed\", \"C\": \"Poor\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000475", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Moderate\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000476", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Moderate\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000476", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Thunderstorm\", \"C\": \"Foggy\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000477", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Thunderstorm\", \"C\": \"Foggy\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000477", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in this scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Not visible\", \"C\": \"Flashing red\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000478", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in this scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Not visible\", \"C\": \"Flashing red\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000478", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state is the road surface in? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Wet\", \"C\": \"Dry\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000479", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/445.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state is the road surface in? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Wet\", \"C\": \"Dry\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000479", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000480", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000480", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dusk\", \"C\": \"Flashing lights\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000481", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dusk\", \"C\": \"Flashing lights\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000481", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000482", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000482", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Foggy\", \"C\": \"Rainy\", \"D\": \"Snowstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000483", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Foggy\", \"C\": \"Rainy\", \"D\": \"Snowstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000483", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the likely state of traffic signals? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Red\", \"C\": \"Flashing yellow\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000484", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the likely state of traffic signals? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Red\", \"C\": \"Flashing yellow\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000484", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Smooth\", \"C\": \"Icy\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000485", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/454.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Smooth\", \"C\": \"Icy\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000485", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Roundabout\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000486", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Roundabout\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000486", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Low\", \"C\": \"Dim streetlight visibility\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000487", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Low\", \"C\": \"Dim streetlight visibility\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000487", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Lane merging\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000488", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Lane merging\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000488", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Stormy\", \"C\": \"Rain\", \"D\": \"Windy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000489", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Stormy\", \"C\": \"Rain\", \"D\": \"Windy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rain" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000489", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the likely state of the traffic signal? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Green\", \"C\": \"Flashing red\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000490", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the likely state of the traffic signal? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Green\", \"C\": \"Flashing red\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000490", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Rough surface\", \"C\": \"Water puddles\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000491", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/463.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Rough surface\", \"C\": \"Water puddles\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000491", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Airport roadway\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000492", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Airport roadway\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000492", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Overexposed\", \"C\": \"Low\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000493", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Overexposed\", \"C\": \"Low\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000493", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Lane merging\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000494", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Lane merging\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000494", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Rainy\", \"C\": \"Dusty\", \"D\": \"Strong wind\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000495", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Rainy\", \"C\": \"Dusty\", \"D\": \"Strong wind\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000495", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state are the traffic signals in? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Green\", \"C\": \"Countdown timer visible\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000496", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state are the traffic signals in? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Green\", \"C\": \"Countdown timer visible\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000496", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Smooth\", \"C\": \"Slippery\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000497", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/472.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Smooth\", \"C\": \"Slippery\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000497", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Rural\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000498", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Rural\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000498", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Partly cloudy\", \"C\": \"Dawn\", \"D\": \"Reflected glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000499", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Partly cloudy\", \"C\": \"Dawn\", \"D\": \"Reflected glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000499", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000500", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000500", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Rainy\", \"C\": \"Gloomy\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000501", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Rainy\", \"C\": \"Gloomy\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000501", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Blinking green\", \"C\": \"No\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000502", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Blinking green\", \"C\": \"No\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000502", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Potholes\", \"C\": \"Smooth\", \"D\": \"Cracked asphalt\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000503", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/481.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Potholes\", \"C\": \"Smooth\", \"D\": \"Cracked asphalt\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000503", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is this? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Shopping district\", \"C\": \"Suburban\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000504", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is this? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Shopping district\", \"C\": \"Suburban\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000504", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this driving scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low\", \"C\": \"Low light\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000505", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this driving scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low\", \"C\": \"Low light\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000505", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000506", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000506", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Mist\", \"C\": \"Clear\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000507", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Mist\", \"C\": \"Clear\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000507", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Flashing red\", \"C\": \"Blinking green\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000508", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Flashing red\", \"C\": \"Blinking green\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000508", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Slippery\", \"C\": \"Oil slick\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000509", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/490.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Slippery\", \"C\": \"Oil slick\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000509", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Service station\", \"C\": \"Market street\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000510", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Service station\", \"C\": \"Market street\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000510", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Partly cloudy\", \"C\": \"Blinding sunlight\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000511", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Partly cloudy\", \"C\": \"Blinding sunlight\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000511", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000512", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000512", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Snowstorm\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000513", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Snowstorm\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000513", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state can be inferred for the traffic signal? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Reflected signal\", \"C\": \"Green\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000514", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state can be inferred for the traffic signal? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Reflected signal\", \"C\": \"Green\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000514", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Construction debris\", \"C\": \"Oil slick\", \"D\": \"Uneven\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000515", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/5.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Construction debris\", \"C\": \"Oil slick\", \"D\": \"Uneven\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000515", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Urban\", \"C\": \"Market street\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000516", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Urban\", \"C\": \"Market street\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000516", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Low\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000517", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Low\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000517", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Lane merging\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000518", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Lane merging\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000518", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Overcast\", \"C\": \"Sleet\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000519", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Overcast\", \"C\": \"Sleet\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000519", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the likely state of the traffic signal? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Flashing red\", \"C\": \"Green\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000520", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the likely state of the traffic signal? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Flashing red\", \"C\": \"Green\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000520", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Snow-covered\", \"C\": \"Cracked asphalt\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000521", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/508.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Snow-covered\", \"C\": \"Cracked asphalt\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000521", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Highway\", \"C\": \"Market street\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000522", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Highway\", \"C\": \"Market street\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000522", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Under tunnel\", \"C\": \"Flashing lights\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000523", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Under tunnel\", \"C\": \"Flashing lights\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000523", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000524", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000524", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Snowstorm\", \"C\": \"Thunderstorm\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000525", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Snowstorm\", \"C\": \"Thunderstorm\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000525", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Blinking green\", \"C\": \"Off / malfunctioning\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000526", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Blinking green\", \"C\": \"Off / malfunctioning\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000526", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Potholes\", \"C\": \"Flooded\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000527", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/517.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Potholes\", \"C\": \"Flooded\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000527", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Parking lot\", \"C\": \"Harbor area\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000528", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Parking lot\", \"C\": \"Harbor area\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000528", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Overexposed\", \"C\": \"Low\", \"D\": \"Reflected glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000529", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Overexposed\", \"C\": \"Low\", \"D\": \"Reflected glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000529", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000530", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000530", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Snowy\", \"C\": \"Foggy\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000531", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Snowy\", \"C\": \"Foggy\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000531", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic signal state likely showing? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Temporary road sign light\", \"C\": \"Yellow\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000532", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic signal state likely showing? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Temporary road sign light\", \"C\": \"Yellow\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000532", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface in this scene? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Smooth\", \"C\": \"Flooded\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000533", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/526.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface in this scene? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Smooth\", \"C\": \"Flooded\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000533", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Shopping district\", \"C\": \"Urban\", \"D\": \"Campus area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000534", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Shopping district\", \"C\": \"Urban\", \"D\": \"Campus area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000534", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Partly cloudy\", \"C\": \"Low light\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000535", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Partly cloudy\", \"C\": \"Low light\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000535", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000536", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000536", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Clear\", \"C\": \"Gloomy\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000537", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Clear\", \"C\": \"Gloomy\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000537", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Temporary road sign light\", \"C\": \"Yellow\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000538", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Temporary road sign light\", \"C\": \"Yellow\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000538", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Wet\", \"C\": \"Muddy\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000539", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/535.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Wet\", \"C\": \"Muddy\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000539", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Urban\", \"C\": \"Shopping district\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000540", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Urban\", \"C\": \"Shopping district\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000540", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Low\", \"C\": \"Foggy\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000541", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Low\", \"C\": \"Foggy\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000541", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000542", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000542", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Sleet\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000543", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Sleet\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000543", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Green\", \"C\": \"Dim / unclear\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000544", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Green\", \"C\": \"Dim / unclear\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000544", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Water puddles\", \"C\": \"Blocked by obstacle\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000545", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/544.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Water puddles\", \"C\": \"Blocked by obstacle\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000545", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this drive in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Market street\", \"C\": \"Campus area\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000546", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this drive in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Market street\", \"C\": \"Campus area\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000546", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Blinding sunlight\", \"C\": \"Low light\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000547", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Blinding sunlight\", \"C\": \"Low light\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000547", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000548", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000548", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Drizzling\", \"C\": \"Overcast\", \"D\": \"Strong wind\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000549", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Drizzling\", \"C\": \"Overcast\", \"D\": \"Strong wind\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000549", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Yellow\", \"C\": \"Reflected signal\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000550", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Yellow\", \"C\": \"Reflected signal\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000550", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000551", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/553.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000551", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Bridge\", \"C\": \"Harbor area\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000552", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Bridge\", \"C\": \"Harbor area\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000552", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Flashing lights\", \"C\": \"Low\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000553", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Flashing lights\", \"C\": \"Low\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000553", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are currently moving ahead? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Moderate\", \"C\": \"Two\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000554", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are currently moving ahead? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Moderate\", \"C\": \"Two\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Two" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000554", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Snowy\", \"C\": \"Foggy\", \"D\": \"Rain\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000555", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Snowy\", \"C\": \"Foggy\", \"D\": \"Rain\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rain" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000555", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"No\", \"C\": \"Reflected signal\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000556", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"No\", \"C\": \"Reflected signal\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000556", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Leaf-covered\", \"C\": \"Gravel\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000557", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/562.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Leaf-covered\", \"C\": \"Gravel\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000557", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Harbor area\", \"C\": \"Parking lot\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000558", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Harbor area\", \"C\": \"Parking lot\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000558", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Smoky\", \"C\": \"Partly cloudy\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000559", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Smoky\", \"C\": \"Partly cloudy\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000559", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000560", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000560", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Flurries\", \"C\": \"Clear\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000561", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Flurries\", \"C\": \"Clear\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000561", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Hidden\", \"C\": \"Countdown timer visible\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000562", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Hidden\", \"C\": \"Countdown timer visible\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000562", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Wet\", \"C\": \"Icy\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000563", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/571.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Wet\", \"C\": \"Icy\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000563", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Subway entrance\", \"C\": \"Shopping district\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000564", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Subway entrance\", \"C\": \"Shopping district\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000564", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Partly cloudy\", \"C\": \"Backlit\", \"D\": \"Reflected glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000565", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Partly cloudy\", \"C\": \"Backlit\", \"D\": \"Reflected glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000565", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000566", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000566", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Rainy\", \"C\": \"Sandstorm\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000567", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Rainy\", \"C\": \"Sandstorm\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000567", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Temporary road sign light\", \"C\": \"Flashing yellow\", \"D\": \"Hidden\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000568", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Temporary road sign light\", \"C\": \"Flashing yellow\", \"D\": \"Hidden\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000568", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Wet\", \"C\": \"Rough surface\", \"D\": \"Cracked asphalt\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000569", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/580.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Wet\", \"C\": \"Rough surface\", \"D\": \"Cracked asphalt\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000569", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Campus area\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000570", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Campus area\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000570", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Dawn\", \"C\": \"Blinding sunlight\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000571", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Dawn\", \"C\": \"Blinding sunlight\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000571", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000572", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000572", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any precipitation or fog present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Clear\", \"C\": \"Rainy\", \"D\": \"Overcast\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000573", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any precipitation or fog present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Clear\", \"C\": \"Rainy\", \"D\": \"Overcast\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000573", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state do the traffic signals appear to be in? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Countdown timer visible\", \"C\": \"Dim / unclear\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000574", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state do the traffic signals appear to be in? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Countdown timer visible\", \"C\": \"Dim / unclear\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000574", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Slippery\", \"C\": \"Icy\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000575", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/59.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Slippery\", \"C\": \"Icy\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000575", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Service station\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000576", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Service station\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000576", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Hazy\", \"C\": \"Strong shadows\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000577", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Hazy\", \"C\": \"Strong shadows\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000577", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic situation like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Accident blockage\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000578", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic situation like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Accident blockage\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000578", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Rainy\", \"C\": \"Dusty\", \"D\": \"Mist\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000579", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Rainy\", \"C\": \"Dusty\", \"D\": \"Mist\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000579", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Countdown timer visible\", \"C\": \"Blinking green\", \"D\": \"No\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000580", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Countdown timer visible\", \"C\": \"Blinking green\", \"D\": \"No\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000580", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Muddy\", \"C\": \"Slippery\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000581", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/599.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Muddy\", \"C\": \"Slippery\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000581", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Construction zone\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000582", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Construction zone\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000582", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Poor\", \"C\": \"Glare\", \"D\": \"Blinding sunlight\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000583", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Poor\", \"C\": \"Glare\", \"D\": \"Blinding sunlight\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000583", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Lane merging\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000584", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Lane merging\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000584", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Mist\", \"C\": \"Rainy\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000585", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Mist\", \"C\": \"Rainy\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000585", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"None\", \"C\": \"Covered by tree\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000586", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"None\", \"C\": \"Covered by tree\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000586", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current state of the road surface? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Snowy\", \"C\": \"Wet\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000587", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/607.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current state of the road surface? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Snowy\", \"C\": \"Wet\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000587", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000588", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000588", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Blinding sunlight\", \"C\": \"Low light\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000589", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Blinding sunlight\", \"C\": \"Low light\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000589", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000590", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000590", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Drizzling\", \"C\": \"Hailing\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000591", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Drizzling\", \"C\": \"Hailing\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000591", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Pedestrian signal\", \"C\": \"Covered by tree\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000592", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Pedestrian signal\", \"C\": \"Covered by tree\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000592", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Icy\", \"C\": \"Wet\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000593", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/616.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Icy\", \"C\": \"Wet\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000593", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Urban\", \"C\": \"Airport roadway\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000594", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Urban\", \"C\": \"Airport roadway\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000594", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Partly cloudy\", \"C\": \"Strong shadows\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000595", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Partly cloudy\", \"C\": \"Strong shadows\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000595", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000596", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000596", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Snowy\", \"C\": \"Sunny\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000597", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Snowy\", \"C\": \"Sunny\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000597", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Countdown timer visible\", \"C\": \"Green\", \"D\": \"Hidden\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000598", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Countdown timer visible\", \"C\": \"Green\", \"D\": \"Hidden\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000598", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Icy\", \"C\": \"Wet\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000599", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/625.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Icy\", \"C\": \"Wet\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000599", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Service station\", \"C\": \"Construction zone\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000600", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Service station\", \"C\": \"Construction zone\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000600", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Hazy\", \"C\": \"Glare\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000601", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Hazy\", \"C\": \"Glare\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000601", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000602", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000602", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Overcast\", \"C\": \"Drizzling\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000603", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Overcast\", \"C\": \"Drizzling\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rainy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000603", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Unknown\", \"C\": \"Yellow\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000604", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Unknown\", \"C\": \"Yellow\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000604", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Leaf-covered\", \"C\": \"Cracked asphalt\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000605", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/634.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Leaf-covered\", \"C\": \"Cracked asphalt\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000605", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Urban\", \"C\": \"Bridge\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000606", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Urban\", \"C\": \"Bridge\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000606", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Under tunnel\", \"C\": \"Partly cloudy\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000607", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Under tunnel\", \"C\": \"Partly cloudy\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000607", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic congestion? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Moderate\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000608", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic congestion? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Moderate\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000608", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any precipitation or haze in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hailing\", \"C\": \"Stormy\", \"D\": \"Strong wind\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000609", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any precipitation or haze in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hailing\", \"C\": \"Stormy\", \"D\": \"Strong wind\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000609", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in this scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Not visible\", \"C\": \"Flashing amber\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000610", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in this scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Not visible\", \"C\": \"Flashing amber\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000610", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Sandy\", \"C\": \"Cracked asphalt\", \"D\": \"Potholes\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000611", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/643.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Sandy\", \"C\": \"Cracked asphalt\", \"D\": \"Potholes\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000611", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Bridge\", \"C\": \"Rural\", \"D\": \"Residential\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000612", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Bridge\", \"C\": \"Rural\", \"D\": \"Residential\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000612", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Foggy\", \"C\": \"Strong shadows\", \"D\": \"Blinding sunlight\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000613", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Foggy\", \"C\": \"Strong shadows\", \"D\": \"Blinding sunlight\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000613", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000614", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000614", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Snowy\", \"C\": \"Strong wind\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000615", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Snowy\", \"C\": \"Strong wind\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000615", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Blinking green\", \"C\": \"Reflected signal\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000616", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Blinking green\", \"C\": \"Reflected signal\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000616", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Dry\", \"C\": \"Smooth\", \"D\": \"Gravel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000617", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/652.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Dry\", \"C\": \"Smooth\", \"D\": \"Gravel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000617", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Urban\", \"C\": \"Harbor area\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000618", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Urban\", \"C\": \"Harbor area\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000618", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Smoky\", \"C\": \"Poor\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000619", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Smoky\", \"C\": \"Poor\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000619", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000620", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000620", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Snowy\", \"C\": \"Drizzling\", \"D\": \"Sleet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000621", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Snowy\", \"C\": \"Drizzling\", \"D\": \"Sleet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000621", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Not visible\", \"C\": \"Green\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000622", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Not visible\", \"C\": \"Green\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000622", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Dry\", \"C\": \"Blocked by obstacle\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000623", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/661.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Dry\", \"C\": \"Blocked by obstacle\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000623", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Shopping district\", \"C\": \"Market street\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000624", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Shopping district\", \"C\": \"Market street\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000624", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Partly cloudy\", \"C\": \"Flashing lights\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000625", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Partly cloudy\", \"C\": \"Flashing lights\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000625", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are present in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000626", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are present in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000626", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is precipitation currently occurring? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Smoggy\", \"C\": \"No\", \"D\": \"Snowstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000627", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is precipitation currently occurring? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Smoggy\", \"C\": \"No\", \"D\": \"Snowstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000627", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Off / malfunctioning\", \"C\": \"Unknown\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000628", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Off / malfunctioning\", \"C\": \"Unknown\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000628", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Dry\", \"C\": \"Bumpy\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000629", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/670.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Dry\", \"C\": \"Bumpy\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000629", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Urban\", \"C\": \"Rural\", \"D\": \"Construction zone\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000630", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Urban\", \"C\": \"Rural\", \"D\": \"Construction zone\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000630", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Backlit\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000631", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Backlit\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000631", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000632", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000632", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Hailing\", \"C\": \"Sunny\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000633", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Hailing\", \"C\": \"Sunny\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000633", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Unknown\", \"C\": \"Temporary road sign light\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000634", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Unknown\", \"C\": \"Temporary road sign light\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000634", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Smooth\", \"C\": \"Icy\", \"D\": \"Cracked asphalt\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000635", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/68.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Smooth\", \"C\": \"Icy\", \"D\": \"Cracked asphalt\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000635", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000636", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000636", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Strong shadows\", \"C\": \"Flashing lights\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000637", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Strong shadows\", \"C\": \"Flashing lights\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000637", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000638", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000638", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Clear\", \"C\": \"Snowy\", \"D\": \"Windy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000639", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Clear\", \"C\": \"Snowy\", \"D\": \"Windy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000639", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Not visible\", \"C\": \"Red\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000640", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Not visible\", \"C\": \"Red\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000640", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snow-covered\", \"C\": \"Gravel\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000641", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/689.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snow-covered\", \"C\": \"Gravel\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000641", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Shopping district\", \"C\": \"Construction zone\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000642", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Shopping district\", \"C\": \"Construction zone\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000642", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Low light\", \"D\": \"Under tunnel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000643", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Low light\", \"D\": \"Under tunnel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000643", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000644", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000644", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hailing\", \"C\": \"Overcast\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000645", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hailing\", \"C\": \"Overcast\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000645", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Hidden\", \"C\": \"Blinking green\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000646", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Hidden\", \"C\": \"Blinking green\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000646", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000647", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/698.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000647", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Shopping district\", \"C\": \"Construction zone\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000648", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Shopping district\", \"C\": \"Construction zone\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000648", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Poor\", \"C\": \"Blinding sunlight\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000649", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Poor\", \"C\": \"Blinding sunlight\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000649", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in motion? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Two\", \"C\": \"Lane merging\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000650", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in motion? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Two\", \"C\": \"Lane merging\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Two" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000650", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Rainy\", \"C\": \"Snowy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000651", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Rainy\", \"C\": \"Snowy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000651", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Reflected signal\", \"C\": \"No\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000652", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Reflected signal\", \"C\": \"No\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000652", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Potholes\", \"C\": \"Dry\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000653", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/706.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Potholes\", \"C\": \"Dry\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000653", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Urban\", \"C\": \"Market street\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000654", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Urban\", \"C\": \"Market street\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000654", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Low\", \"C\": \"Hazy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000655", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Low\", \"C\": \"Hazy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000655", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000656", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000656", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is likely present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Rainy\", \"C\": \"Windy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000657", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is likely present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Rainy\", \"C\": \"Windy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000657", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"No\", \"C\": \"Blinking green\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000658", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"No\", \"C\": \"Blinking green\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000658", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Gravel\", \"C\": \"Newly paved\", \"D\": \"Leaf-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000659", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/715.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Gravel\", \"C\": \"Newly paved\", \"D\": \"Leaf-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000659", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Highway\", \"C\": \"Parking lot\", \"D\": \"Residential avenue\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000660", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Highway\", \"C\": \"Parking lot\", \"D\": \"Residential avenue\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential avenue" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000660", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Overexposed\", \"C\": \"Low light\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000661", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Overexposed\", \"C\": \"Low light\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000661", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000662", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000662", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Windy\", \"C\": \"Clear\", \"D\": \"Sleet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000663", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Windy\", \"C\": \"Clear\", \"D\": \"Sleet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000663", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Flashing red\", \"C\": \"Temporary road sign light\", \"D\": \"No\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000664", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Flashing red\", \"C\": \"Temporary road sign light\", \"D\": \"No\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000664", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Leaf-covered\", \"C\": \"Wet\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000665", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/724.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Leaf-covered\", \"C\": \"Wet\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000665", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Construction zone\", \"C\": \"Market street\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000666", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Construction zone\", \"C\": \"Market street\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000666", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Poor\", \"C\": \"Strong shadows\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000667", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Poor\", \"C\": \"Strong shadows\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000667", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Lane merging\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000668", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Lane merging\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000668", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Sunny\", \"C\": \"Fog\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000669", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Sunny\", \"C\": \"Fog\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Fog" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000669", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Missing\", \"C\": \"Temporary road sign light\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000670", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Missing\", \"C\": \"Temporary road sign light\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Missing" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000670", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Dry\", \"C\": \"Slippery\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000671", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/733.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Dry\", \"C\": \"Slippery\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000671", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Parkway\", \"C\": \"Airport roadway\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000672", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Parkway\", \"C\": \"Airport roadway\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Parkway" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000672", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Hazy\", \"C\": \"Poor\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000673", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Hazy\", \"C\": \"Poor\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000673", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000674", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000674", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Snowstorm\", \"C\": \"Clear\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000675", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Snowstorm\", \"C\": \"Clear\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000675", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Covered by tree\", \"C\": \"Reflected signal\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000676", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Covered by tree\", \"C\": \"Reflected signal\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000676", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Blocked by obstacle\", \"C\": \"Cracked asphalt\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000677", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/742.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Blocked by obstacle\", \"C\": \"Cracked asphalt\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000677", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Highway\", \"C\": \"Service station\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000678", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Highway\", \"C\": \"Service station\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000678", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Rain-blurred\", \"C\": \"Poor\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000679", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Rain-blurred\", \"C\": \"Poor\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000679", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible moving? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"One\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000680", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible moving? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"One\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000680", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Smoggy\", \"C\": \"Dusty\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000681", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Smoggy\", \"C\": \"Dusty\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000681", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Unknown\", \"C\": \"Off / malfunctioning\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000682", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Unknown\", \"C\": \"Off / malfunctioning\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000682", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Slippery\", \"C\": \"Wet\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000683", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/751.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Slippery\", \"C\": \"Wet\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000683", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000684", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000684", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Dusk\", \"C\": \"Low\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000685", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Dusk\", \"C\": \"Low\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000685", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic volume high or light? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000686", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic volume high or light? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000686", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Mist\", \"C\": \"Clear\", \"D\": \"Drizzling\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000687", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Mist\", \"C\": \"Clear\", \"D\": \"Drizzling\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000687", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Unknown\", \"C\": \"Off / malfunctioning\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000688", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Unknown\", \"C\": \"Off / malfunctioning\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000688", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Newly paved\", \"C\": \"Flooded\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000689", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/760.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Newly paved\", \"C\": \"Flooded\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000689", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Highway\", \"C\": \"Construction zone\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000690", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Highway\", \"C\": \"Construction zone\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Highway" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000690", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Low\", \"C\": \"Reflected glare\", \"D\": \"Blinding sunlight\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000691", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Low\", \"C\": \"Reflected glare\", \"D\": \"Blinding sunlight\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000691", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Lane merging\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000692", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Lane merging\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000692", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Clear\", \"C\": \"Overcast\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000693", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Clear\", \"C\": \"Overcast\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000693", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state is the traffic signal likely in? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Flashing red\", \"C\": \"Reflected signal\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000694", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state is the traffic signal likely in? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Flashing red\", \"C\": \"Reflected signal\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000694", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Dry\", \"C\": \"Gravel\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000695", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/77.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Dry\", \"C\": \"Gravel\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000695", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Highway\", \"C\": \"Forest road\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000696", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Highway\", \"C\": \"Forest road\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000696", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Low light\", \"C\": \"Clear\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000697", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Low light\", \"C\": \"Clear\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000697", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the traffic? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"One\", \"C\": \"Heavy\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000698", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the traffic? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"One\", \"C\": \"Heavy\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000698", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Clear\", \"C\": \"Sunny\", \"D\": \"Dusty\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000699", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Clear\", \"C\": \"Sunny\", \"D\": \"Dusty\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000699", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Covered by tree\", \"C\": \"Absent\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000700", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Covered by tree\", \"C\": \"Absent\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000700", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Sandy\", \"C\": \"Muddy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000701", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/779.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Sandy\", \"C\": \"Muddy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000701", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Two-lane road\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000702", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Two-lane road\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Two-lane road" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000702", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Under tunnel\", \"C\": \"Hazy\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000703", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Under tunnel\", \"C\": \"Hazy\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000703", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic flow like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000704", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic flow like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000704", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Clear\", \"C\": \"Sandstorm\", \"D\": \"Overcast\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000705", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Clear\", \"C\": \"Sandstorm\", \"D\": \"Overcast\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000705", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Absent\", \"C\": \"Green\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000706", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Absent\", \"C\": \"Green\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000706", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Dry\", \"C\": \"Leaf-covered\", \"D\": \"Gravel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000707", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/788.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Dry\", \"C\": \"Leaf-covered\", \"D\": \"Gravel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000707", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Parking lot\", \"C\": \"Construction zone\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000708", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Parking lot\", \"C\": \"Construction zone\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000708", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Partly cloudy\", \"C\": \"Hazy\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000709", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Partly cloudy\", \"C\": \"Hazy\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000709", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible ahead? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"One\", \"C\": \"Congested\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000710", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible ahead? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"One\", \"C\": \"Congested\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000710", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Clear\", \"C\": \"Drizzling\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000711", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Clear\", \"C\": \"Drizzling\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000711", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Flashing yellow\", \"C\": \"Unknown\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000712", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Flashing yellow\", \"C\": \"Unknown\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000712", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Partially damaged\", \"C\": \"Snowy\", \"D\": \"Debris scattered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000713", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/797.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Partially damaged\", \"C\": \"Snowy\", \"D\": \"Debris scattered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000713", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Rural\", \"C\": \"Urban\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000714", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Rural\", \"C\": \"Urban\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000714", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Hazy\", \"C\": \"Foggy\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000715", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Hazy\", \"C\": \"Foggy\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000715", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Accident blockage\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000716", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Accident blockage\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000716", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Strong wind\", \"C\": \"Hailing\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000717", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Strong wind\", \"C\": \"Hailing\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000717", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"No\", \"C\": \"Temporary road sign light\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000718", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"No\", \"C\": \"Temporary road sign light\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000718", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Icy\", \"C\": \"Dry\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000719", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/805.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Icy\", \"C\": \"Dry\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000719", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Rural\", \"C\": \"Highway\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000720", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Rural\", \"C\": \"Highway\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000720", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Rain-blurred\", \"C\": \"Poor\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000721", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Rain-blurred\", \"C\": \"Poor\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000721", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible on the road? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Congested\", \"D\": \"One\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000722", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible on the road? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Congested\", \"D\": \"One\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000722", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Clear\", \"C\": \"Hailing\", \"D\": \"Sleet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000723", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Clear\", \"C\": \"Hailing\", \"D\": \"Sleet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000723", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Flashing yellow\", \"C\": \"No\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000724", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Flashing yellow\", \"C\": \"No\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000724", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Dry\", \"C\": \"Partially damaged\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000725", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/814.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Dry\", \"C\": \"Partially damaged\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000725", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this scene set in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Market street\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000726", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this scene set in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Market street\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000726", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Low light\", \"C\": \"Low\", \"D\": \"Blinding sunlight\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000727", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Low light\", \"C\": \"Low\", \"D\": \"Blinding sunlight\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000727", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Lane merging\", \"C\": \"One\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000728", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Lane merging\", \"C\": \"One\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000728", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Cloudy\", \"C\": \"Clear\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000729", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Cloudy\", \"C\": \"Clear\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000729", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Red\", \"C\": \"Unknown\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000730", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Red\", \"C\": \"Unknown\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000730", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Slippery\", \"B\": \"Dry\", \"C\": \"Uneven\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000731", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/823.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Slippery\", \"B\": \"Dry\", \"C\": \"Uneven\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000731", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the image? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Urban\", \"C\": \"Forest road\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000732", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the image? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Urban\", \"C\": \"Forest road\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000732", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Rain-blurred\", \"C\": \"Low light\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000733", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Rain-blurred\", \"C\": \"Low light\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000733", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000734", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000734", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Smoggy\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000735", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Smoggy\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000735", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the image? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing amber\", \"C\": \"Blinking green\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000736", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the image? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing amber\", \"C\": \"Blinking green\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000736", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Construction debris\", \"C\": \"Sandy\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000737", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/832.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Construction debris\", \"C\": \"Sandy\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000737", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Subway entrance\", \"C\": \"Two-lane street\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000738", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Subway entrance\", \"C\": \"Two-lane street\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Two-lane street" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000738", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Clear\", \"C\": \"Smoky\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000739", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Clear\", \"C\": \"Smoky\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000739", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What can be inferred about vehicle movement? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Lane merging\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000740", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What can be inferred about vehicle movement? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Lane merging\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000740", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there evidence of precipitation or fog? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Thunderstorm\", \"C\": \"Stormy\", \"D\": \"Windy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000741", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there evidence of precipitation or fog? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Thunderstorm\", \"C\": \"Stormy\", \"D\": \"Windy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000741", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing yellow\", \"C\": \"Flashing red\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000742", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing yellow\", \"C\": \"Flashing red\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000742", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Newly paved\", \"C\": \"Dry\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000743", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/841.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Newly paved\", \"C\": \"Dry\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000743", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Coastal road\", \"C\": \"Suburban\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000744", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Coastal road\", \"C\": \"Suburban\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000744", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Dawn\", \"C\": \"Poor\", \"D\": \"Glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000745", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Dawn\", \"C\": \"Poor\", \"D\": \"Glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000745", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000746", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000746", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Stormy\", \"C\": \"Gloomy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000747", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Stormy\", \"C\": \"Gloomy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000747", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state might the traffic light ahead show? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Reflected signal\", \"C\": \"Green\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000748", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state might the traffic light ahead show? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Reflected signal\", \"C\": \"Green\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000748", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Gravel\", \"C\": \"Dry\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000749", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/850.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Gravel\", \"C\": \"Dry\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000749", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Airport roadway\", \"C\": \"Rural\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000750", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Airport roadway\", \"C\": \"Rural\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000750", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Low light\", \"C\": \"Low\", \"D\": \"Under tunnel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000751", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Low light\", \"C\": \"Low\", \"D\": \"Under tunnel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000751", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000752", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic movement? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000752", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Clear\", \"C\": \"Flurries\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000753", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Clear\", \"C\": \"Flurries\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000753", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the likely state of the traffic signal? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Flashing amber\", \"C\": \"Green\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000754", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the likely state of the traffic signal? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Flashing amber\", \"C\": \"Green\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000754", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Leaf-covered\", \"C\": \"Snow-covered\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000755", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/86.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Leaf-covered\", \"C\": \"Snow-covered\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000755", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000756", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000756", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Flashing lights\", \"C\": \"Dawn\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000757", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Flashing lights\", \"C\": \"Dawn\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000757", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the traffic? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"One\", \"C\": \"Moderate\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000758", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the traffic? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"One\", \"C\": \"Moderate\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000758", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dusty\", \"C\": \"Flurries\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000759", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dusty\", \"C\": \"Flurries\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000759", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light signal be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Red\", \"C\": \"Dim / unclear\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000760", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light signal be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Red\", \"C\": \"Dim / unclear\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000760", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Flooded\", \"C\": \"Dry\", \"D\": \"Leaf-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000761", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/869.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Flooded\", \"C\": \"Dry\", \"D\": \"Leaf-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000761", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Urban\", \"C\": \"Campus area\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000762", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Urban\", \"C\": \"Campus area\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000762", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Flashing lights\", \"C\": \"Poor\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000763", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Flashing lights\", \"C\": \"Poor\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000763", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible on the road? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Heavy\", \"D\": \"One\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000764", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible on the road? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Heavy\", \"D\": \"One\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "One" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000764", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Mist\", \"C\": \"Stormy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000765", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Mist\", \"C\": \"Stormy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000765", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing amber\", \"C\": \"Green\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000766", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing amber\", \"C\": \"Green\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000766", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Smooth\", \"C\": \"Newly paved\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000767", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/878.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Smooth\", \"C\": \"Newly paved\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000767", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Roundabout\", \"C\": \"Campus area\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000768", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Roundabout\", \"C\": \"Campus area\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000768", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this driving scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Blinding sunlight\", \"C\": \"Overexposed\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000769", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this driving scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Blinding sunlight\", \"C\": \"Overexposed\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000769", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000770", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000770", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Clear\", \"C\": \"Dusty\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000771", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Clear\", \"C\": \"Dusty\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000771", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Green\", \"C\": \"Unknown\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000772", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Green\", \"C\": \"Unknown\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000772", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Sandy\", \"C\": \"Blocked by obstacle\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000773", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/887.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Sandy\", \"C\": \"Blocked by obstacle\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000773", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Parking lot\", \"C\": \"Urban\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000774", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Parking lot\", \"C\": \"Urban\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000774", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Backlit\", \"C\": \"Flashing lights\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000775", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low\", \"B\": \"Backlit\", \"C\": \"Flashing lights\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000775", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000776", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000776", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any precipitation present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Clear\", \"C\": \"Rainy\", \"D\": \"Sleet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000777", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any precipitation present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Clear\", \"C\": \"Rainy\", \"D\": \"Sleet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000777", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Unknown\", \"C\": \"Dim / unclear\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000778", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Unknown\", \"C\": \"Dim / unclear\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000778", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Flooded\", \"C\": \"Dry\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000779", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/896.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Flooded\", \"C\": \"Dry\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000779", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Shopping district\", \"C\": \"Construction zone\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000780", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Shopping district\", \"C\": \"Construction zone\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000780", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Partly cloudy\", \"C\": \"Poor\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000781", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Partly cloudy\", \"C\": \"Poor\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000781", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are currently on the road? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Few\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000782", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are currently on the road? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Few\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Few" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000782", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Clear\", \"C\": \"Hailing\", \"D\": \"Strong wind\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000783", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Clear\", \"C\": \"Hailing\", \"D\": \"Strong wind\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000783", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Flashing red\", \"C\": \"Covered by tree\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000784", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Flashing red\", \"C\": \"Covered by tree\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000784", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rough surface\", \"C\": \"Blocked by obstacle\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000785", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/904.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rough surface\", \"C\": \"Blocked by obstacle\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000785", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Coastal road\", \"C\": \"Market street\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000786", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Coastal road\", \"C\": \"Market street\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000786", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Overexposed\", \"C\": \"Glare\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000787", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Overexposed\", \"C\": \"Glare\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000787", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000788", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000788", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Sandstorm\", \"C\": \"Clear\", \"D\": \"Overcast\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000789", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Sandstorm\", \"C\": \"Clear\", \"D\": \"Overcast\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000789", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Missing\", \"C\": \"Yellow\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000790", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Missing\", \"C\": \"Yellow\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Missing" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000790", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Potholes\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000791", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/913.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Potholes\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000791", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Parking lot\", \"C\": \"Urban\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000792", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Parking lot\", \"C\": \"Urban\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000792", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Dim streetlight visibility\", \"C\": \"Low\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000793", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Dim streetlight visibility\", \"C\": \"Low\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000793", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000794", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000794", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Cloudy\", \"C\": \"Smoggy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000795", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Cloudy\", \"C\": \"Smoggy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000795", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Countdown timer visible\", \"C\": \"Unknown\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000796", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Countdown timer visible\", \"C\": \"Unknown\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000796", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000797", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/922.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000797", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Suburban\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000798", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Suburban\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000798", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Low\", \"C\": \"Flashing lights\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000799", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Low\", \"C\": \"Flashing lights\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000799", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000800", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000800", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Clear\", \"C\": \"Sandstorm\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000801", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Clear\", \"C\": \"Sandstorm\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000801", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing red\", \"C\": \"Pedestrian signal\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000802", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing red\", \"C\": \"Pedestrian signal\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000802", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Debris scattered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000803", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/931.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Debris scattered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000803", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Harbor area\", \"C\": \"Construction zone\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000804", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Harbor area\", \"C\": \"Construction zone\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000804", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Dim streetlight visibility\", \"C\": \"Dawn\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000805", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Dim streetlight visibility\", \"C\": \"Dawn\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000805", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000806", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000806", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000807", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000807", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Unknown\", \"C\": \"Reflected signal\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000808", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Unknown\", \"C\": \"Reflected signal\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000808", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Dry\", \"C\": \"Cracked asphalt\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000809", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/940.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Dry\", \"C\": \"Cracked asphalt\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000809", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Service station\", \"C\": \"Rural\", \"D\": \"Construction zone\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000810", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Service station\", \"C\": \"Rural\", \"D\": \"Construction zone\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000810", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Strong shadows\", \"C\": \"Poor\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000811", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Strong shadows\", \"C\": \"Poor\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000811", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000812", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000812", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Cloudy\", \"C\": \"Smoggy\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000813", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Cloudy\", \"C\": \"Smoggy\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000813", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state is the traffic light likely in? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Dim / unclear\", \"C\": \"Green\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000814", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What state is the traffic light likely in? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Dim / unclear\", \"C\": \"Green\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Green" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000814", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000815", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/95.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000815", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Urban\", \"C\": \"Shopping district\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000816", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Urban\", \"C\": \"Shopping district\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000816", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Partly cloudy\", \"C\": \"Poor\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000817", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Partly cloudy\", \"C\": \"Poor\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000817", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000818", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000818", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Hailing\", \"C\": \"Dusty\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000819", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Hailing\", \"C\": \"Dusty\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000819", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Green\", \"C\": \"Flashing amber\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000820", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Green\", \"C\": \"Flashing amber\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000820", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Icy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000821", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/959.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Icy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000821", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Rural\", \"C\": \"Service station\", \"D\": \"Construction zone\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000822", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Rural\", \"C\": \"Service station\", \"D\": \"Construction zone\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000822", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Reflected glare\", \"C\": \"Strong shadows\", \"D\": \"Poor\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000823", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Reflected glare\", \"C\": \"Strong shadows\", \"D\": \"Poor\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000823", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity ahead? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000824", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity ahead? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000824", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Cloudy\", \"C\": \"Snowy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000825", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Cloudy\", \"C\": \"Snowy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000825", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Red\", \"C\": \"Flashing red\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000826", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Red\", \"C\": \"Flashing red\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000826", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Snowy\", \"C\": \"Newly paved\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000827", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/968.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Snowy\", \"C\": \"Newly paved\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000827", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Parking lot\", \"C\": \"Rural\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000828", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Parking lot\", \"C\": \"Rural\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000828", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Low\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000829", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Low\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000829", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there traffic present in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Rush hour\", \"C\": \"Detour traffic\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000830", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there traffic present in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Rush hour\", \"C\": \"Detour traffic\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000830", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Clear\", \"C\": \"Strong wind\", \"D\": \"Windy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000831", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Clear\", \"C\": \"Strong wind\", \"D\": \"Windy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000831", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Yellow\", \"C\": \"Not visible\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000832", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Yellow\", \"C\": \"Not visible\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000832", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Icy\", \"C\": \"Cracked asphalt\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000833", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/977.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Icy\", \"C\": \"Cracked asphalt\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000833", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Bridge\", \"C\": \"Roundabout\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000834", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Bridge\", \"C\": \"Roundabout\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000834", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Low light\", \"C\": \"Backlit\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000835", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Low light\", \"C\": \"Backlit\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000835", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there traffic present in the scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000836", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there traffic present in the scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000836", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Clear\", \"C\": \"Windy\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000837", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Clear\", \"C\": \"Windy\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000837", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Covered by tree\", \"C\": \"Green\", \"D\": \"Missing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000838", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Covered by tree\", \"C\": \"Green\", \"D\": \"Missing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Missing" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000838", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Icy\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000839", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/986.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Icy\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000839", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Urban\", \"C\": \"Two-lane\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000840", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is being driven on? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Urban\", \"C\": \"Two-lane\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Two-lane" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000840", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Dawn\", \"C\": \"Strong shadows\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000841", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Dawn\", \"C\": \"Strong shadows\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000841", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any visible traffic nearby? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Rush hour\", \"C\": \"None\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000842", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any visible traffic nearby? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Rush hour\", \"C\": \"None\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000842", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Overcast\", \"C\": \"Rainy\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000843", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Overcast\", \"C\": \"Rainy\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000843", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Flashing yellow\", \"C\": \"Not visible\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000844", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Flashing yellow\", \"C\": \"Not visible\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000844", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What can be inferred about the road surface? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Partially damaged\", \"C\": \"Smooth\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000845", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.png", "event": "data/perception/pku/aps_frames_sampled/val/low_light/006_val_low_light/995.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What can be inferred about the road surface? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Partially damaged\", \"C\": \"Smooth\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000845", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Urban\", \"C\": \"Service station\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000846", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Urban\", \"C\": \"Service station\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000846", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Overexposed\", \"C\": \"Dawn\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000847", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Overexposed\", \"C\": \"Dawn\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000847", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000848", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000848", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Drizzling\", \"C\": \"Sandstorm\", \"D\": \"Haze\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000849", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Drizzling\", \"C\": \"Sandstorm\", \"D\": \"Haze\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000849", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this frame? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Temporary road sign light\", \"C\": \"Countdown timer visible\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000850", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this frame? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Temporary road sign light\", \"C\": \"Countdown timer visible\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000850", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Leaf-covered\", \"C\": \"Potholes\", \"D\": \"Cracked asphalt\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000851", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/0.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Leaf-covered\", \"C\": \"Potholes\", \"D\": \"Cracked asphalt\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000851", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Market street\", \"C\": \"Service station\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000852", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Market street\", \"C\": \"Service station\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000852", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dusk\", \"C\": \"Hazy\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000853", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dusk\", \"C\": \"Hazy\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000853", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle activity in the scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000854", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle activity in the scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000854", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Drizzling\", \"C\": \"Snowstorm\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000855", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Drizzling\", \"C\": \"Snowstorm\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000855", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing yellow\", \"C\": \"None\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000856", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing yellow\", \"C\": \"None\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000856", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Oil slick\", \"C\": \"Muddy\", \"D\": \"Gravel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000857", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/107.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Oil slick\", \"C\": \"Muddy\", \"D\": \"Gravel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000857", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Parking lot\", \"C\": \"Construction zone\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000858", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Parking lot\", \"C\": \"Construction zone\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000858", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Reflected glare\", \"C\": \"Flashing lights\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000859", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Reflected glare\", \"C\": \"Flashing lights\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000859", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000860", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000860", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any visible weather affecting the scene? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Foggy\", \"C\": \"Rainy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000861", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any visible weather affecting the scene? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Foggy\", \"C\": \"Rainy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000861", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from this view? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Blinking green\", \"C\": \"Dim / unclear\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000862", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from this view? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Blinking green\", \"C\": \"Dim / unclear\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000862", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Blocked by obstacle\", \"C\": \"Dry\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000863", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/116.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Blocked by obstacle\", \"C\": \"Dry\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000863", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Roundabout\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000864", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Roundabout\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000864", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Reflected glare\", \"C\": \"Hazy\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000865", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Reflected glare\", \"C\": \"Hazy\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000865", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000866", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000866", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Thunderstorm\", \"C\": \"Mist\", \"D\": \"Haze\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000867", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Thunderstorm\", \"C\": \"Mist\", \"D\": \"Haze\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000867", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Green\", \"C\": \"Off / malfunctioning\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000868", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Green\", \"C\": \"Off / malfunctioning\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000868", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Bumpy\", \"C\": \"Flooded\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000869", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/125.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Bumpy\", \"C\": \"Flooded\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000869", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Shopping district\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000870", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Shopping district\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000870", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Strong shadows\", \"C\": \"Smoky\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000871", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Strong shadows\", \"C\": \"Smoky\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000871", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000872", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000872", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Flurries\", \"C\": \"Haze\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000873", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Flurries\", \"C\": \"Haze\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000873", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Not visible\", \"C\": \"Blinking green\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000874", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Not visible\", \"C\": \"Blinking green\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000874", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Water puddles\", \"C\": \"Bumpy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000875", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/134.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Water puddles\", \"C\": \"Bumpy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000875", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Urban\", \"C\": \"Campus area\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000876", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Urban\", \"C\": \"Campus area\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000876", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting quality in the scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Dim streetlight visibility\", \"C\": \"Glare\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000877", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting quality in the scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Dim streetlight visibility\", \"C\": \"Glare\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000877", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000878", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000878", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Strong wind\", \"C\": \"Haze\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000879", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Strong wind\", \"C\": \"Haze\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000879", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Green\", \"C\": \"Not visible\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000880", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Green\", \"C\": \"Not visible\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000880", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Bumpy\", \"C\": \"Dry\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000881", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/143.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Bumpy\", \"C\": \"Dry\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000881", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Shopping district\", \"C\": \"Roundabout\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000882", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Shopping district\", \"C\": \"Roundabout\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000882", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene described? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Reflected glare\", \"C\": \"Foggy\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000883", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene described? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Reflected glare\", \"C\": \"Foggy\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000883", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000884", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000884", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is implied by the scene? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000885", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is implied by the scene? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000885", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Reflected signal\", \"C\": \"Yellow\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000886", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Reflected signal\", \"C\": \"Yellow\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000886", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Potholes\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000887", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/152.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Potholes\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000887", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Parking lot\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000888", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Parking lot\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000888", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Clear\", \"C\": \"Washed out\", \"D\": \"Under tunnel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000889", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Clear\", \"C\": \"Washed out\", \"D\": \"Under tunnel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Washed out" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000889", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000890", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000890", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Dusty\", \"C\": \"Sleet\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000891", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Dusty\", \"C\": \"Sleet\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000891", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Green\", \"C\": \"Not visible\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000892", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Green\", \"C\": \"Not visible\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000892", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000893", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/161.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000893", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Suburban\", \"C\": \"Market street\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000894", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Suburban\", \"C\": \"Market street\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000894", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dawn\", \"C\": \"Blinding sunlight\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000895", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dawn\", \"C\": \"Blinding sunlight\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000895", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000896", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000896", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Stormy\", \"C\": \"Sunny\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000897", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Stormy\", \"C\": \"Sunny\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000897", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Pedestrian signal\", \"C\": \"Not visible\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000898", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Pedestrian signal\", \"C\": \"Not visible\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000898", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Muddy\", \"C\": \"Smooth\", \"D\": \"Icy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000899", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/170.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Muddy\", \"C\": \"Smooth\", \"D\": \"Icy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000899", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Bridge\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000900", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Bridge\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000900", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Smoky\", \"C\": \"Overexposed\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000901", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Smoky\", \"C\": \"Overexposed\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000901", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition ahead? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000902", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition ahead? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000902", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Gloomy\", \"C\": \"Haze\", \"D\": \"Drizzling\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000903", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Gloomy\", \"C\": \"Haze\", \"D\": \"Drizzling\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000903", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Red\", \"C\": \"Yellow\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000904", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Red\", \"C\": \"Yellow\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000904", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Potholes\", \"C\": \"Newly paved\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000905", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/18.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Potholes\", \"C\": \"Newly paved\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000905", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Coastal road\", \"C\": \"Suburban\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000906", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Coastal road\", \"C\": \"Suburban\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000906", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting condition described? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000907", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting condition described? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000907", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000908", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000908", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Sunny\", \"C\": \"Strong wind\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000909", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Sunny\", \"C\": \"Strong wind\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000909", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Temporary road sign light\", \"C\": \"Pedestrian signal\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000910", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Temporary road sign light\", \"C\": \"Pedestrian signal\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000910", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Slippery\", \"C\": \"Snow-covered\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000911", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/189.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Slippery\", \"C\": \"Snow-covered\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000911", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Construction zone\", \"C\": \"Market street\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000912", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Construction zone\", \"C\": \"Market street\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000912", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Clear\", \"C\": \"Partly cloudy\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000913", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Clear\", \"C\": \"Partly cloudy\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000913", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000914", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000914", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Rainy\", \"C\": \"Dry\", \"D\": \"Thunderstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000915", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Rainy\", \"C\": \"Dry\", \"D\": \"Thunderstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000915", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Countdown timer visible\", \"C\": \"Not visible\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000916", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Countdown timer visible\", \"C\": \"Not visible\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000916", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Newly paved\", \"C\": \"Flooded\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000917", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/198.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Newly paved\", \"C\": \"Flooded\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000917", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Highway\", \"C\": \"Rural\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000918", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Highway\", \"C\": \"Rural\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000918", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Flashing lights\", \"C\": \"Hazy\", \"D\": \"Blinding sunlight\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000919", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Flashing lights\", \"C\": \"Hazy\", \"D\": \"Blinding sunlight\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000919", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Lane merging\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000920", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Lane merging\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000920", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Hailing\", \"C\": \"Snowstorm\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000921", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Hailing\", \"C\": \"Snowstorm\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000921", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"None\", \"C\": \"Green\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000922", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"None\", \"C\": \"Green\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000922", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Slippery\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000923", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/206.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Slippery\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000923", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Roundabout\", \"C\": \"Rural\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000924", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Roundabout\", \"C\": \"Rural\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000924", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Dawn\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000925", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Dawn\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000925", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000926", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000926", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Hailing\", \"C\": \"Dry\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000927", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Hailing\", \"C\": \"Dry\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000927", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Not visible\", \"C\": \"Reflected signal\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000928", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Not visible\", \"C\": \"Reflected signal\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000928", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Gravel\", \"C\": \"Smooth\", \"D\": \"Cracked asphalt\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000929", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/215.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Gravel\", \"C\": \"Smooth\", \"D\": \"Cracked asphalt\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000929", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000930", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000930", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Reflected glare\", \"C\": \"Hazy\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000931", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Reflected glare\", \"C\": \"Hazy\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000931", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Lane merging\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000932", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Lane merging\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000932", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Snowstorm\", \"C\": \"Snowy\", \"D\": \"Haze\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000933", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Snowstorm\", \"C\": \"Snowy\", \"D\": \"Haze\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000933", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Off / malfunctioning\", \"C\": \"Not visible\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000934", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Off / malfunctioning\", \"C\": \"Not visible\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000934", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Icy\", \"C\": \"Construction debris\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000935", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/224.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Icy\", \"C\": \"Construction debris\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000935", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Subway entrance\", \"C\": \"Shopping district\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000936", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Subway entrance\", \"C\": \"Shopping district\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000936", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Hazy\", \"C\": \"Low light\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000937", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Hazy\", \"C\": \"Low light\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000937", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000938", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000938", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Rainy\", \"C\": \"Mist\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000939", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Rainy\", \"C\": \"Mist\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000939", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Not visible\", \"C\": \"Green\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000940", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Not visible\", \"C\": \"Green\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000940", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Rough surface\", \"C\": \"Flooded\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000941", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/233.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Rough surface\", \"C\": \"Flooded\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000941", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Forest road\", \"C\": \"Urban\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000942", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Forest road\", \"C\": \"Urban\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000942", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Dusk\", \"C\": \"Smoky\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000943", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Dusk\", \"C\": \"Smoky\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000943", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Accident blockage\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000944", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Accident blockage\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000944", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Snowstorm\", \"C\": \"Windy\", \"D\": \"Drizzling\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000945", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Snowstorm\", \"C\": \"Windy\", \"D\": \"Drizzling\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000945", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"No\", \"B\": \"None\", \"C\": \"Blinking green\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000946", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"No\", \"B\": \"None\", \"C\": \"Blinking green\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000946", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Sandy\", \"C\": \"Icy\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000947", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/242.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Sandy\", \"C\": \"Icy\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000947", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Construction zone\", \"C\": \"Service station\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000948", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Construction zone\", \"C\": \"Service station\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000948", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Hazy\", \"C\": \"Clear\", \"D\": \"Glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000949", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Hazy\", \"C\": \"Clear\", \"D\": \"Glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000949", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000950", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000950", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Flurries\", \"C\": \"Snowy\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000951", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Flurries\", \"C\": \"Snowy\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000951", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"None\", \"C\": \"Off / malfunctioning\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000952", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"None\", \"C\": \"Off / malfunctioning\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000952", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Oil slick\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000953", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/251.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Oil slick\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000953", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Urban\", \"C\": \"Service station\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000954", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Urban\", \"C\": \"Service station\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000954", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Dusk\", \"C\": \"Overexposed\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000955", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Dusk\", \"C\": \"Overexposed\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000955", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000956", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000956", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Snowstorm\", \"C\": \"Strong wind\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000957", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Snowstorm\", \"C\": \"Strong wind\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000957", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Off / malfunctioning\", \"C\": \"Covered by tree\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000958", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Off / malfunctioning\", \"C\": \"Covered by tree\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000958", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Sandy\", \"C\": \"Smooth\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000959", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/260.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Sandy\", \"C\": \"Smooth\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000959", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Urban\", \"C\": \"Harbor area\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000960", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Urban\", \"C\": \"Harbor area\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000960", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Rain-blurred\", \"C\": \"Clear\", \"D\": \"Reflected glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000961", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Rain-blurred\", \"C\": \"Clear\", \"D\": \"Reflected glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000961", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000962", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000962", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Smoggy\", \"C\": \"Snowstorm\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000963", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Smoggy\", \"C\": \"Snowstorm\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000963", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the scene? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Flashing red\", \"C\": \"Dim / unclear\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000964", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the scene? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Flashing red\", \"C\": \"Dim / unclear\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000964", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Slippery\", \"C\": \"Dry\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000965", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/27.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Slippery\", \"C\": \"Dry\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000965", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Service station\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000966", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Service station\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000966", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Strong shadows\", \"C\": \"Golden\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000967", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Strong shadows\", \"C\": \"Golden\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Golden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000967", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000968", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000968", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Stormy\", \"C\": \"Drizzling\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000969", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Stormy\", \"C\": \"Drizzling\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000969", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Pedestrian signal\", \"C\": \"Dim / unclear\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000970", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Pedestrian signal\", \"C\": \"Dim / unclear\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000970", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Bumpy\", \"C\": \"Uneven\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000971", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/279.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Bumpy\", \"C\": \"Uneven\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000971", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Service station\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000972", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Service station\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000972", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Flashing lights\", \"C\": \"Strong shadows\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000973", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Flashing lights\", \"C\": \"Strong shadows\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000973", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like on this road? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000974", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like on this road? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000974", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Sunny\", \"C\": \"Windy\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000975", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Sunny\", \"C\": \"Windy\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000975", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from this image? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Yellow\", \"C\": \"Unknown\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000976", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from this image? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Yellow\", \"C\": \"Unknown\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000976", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface in this scene? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Uneven\", \"C\": \"Oil slick\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000977", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/288.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface in this scene? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Uneven\", \"C\": \"Oil slick\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000977", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000978", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000978", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the image clarity described? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Strong shadows\", \"C\": \"Under tunnel\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000979", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the image clarity described? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Strong shadows\", \"C\": \"Under tunnel\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000979", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000980", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000980", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is implied by the scene? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Sunny\", \"C\": \"Sandstorm\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000981", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is implied by the scene? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Sunny\", \"C\": \"Sandstorm\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000981", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Countdown timer visible\", \"C\": \"Covered by tree\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000982", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Countdown timer visible\", \"C\": \"Covered by tree\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000982", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Slippery\", \"B\": \"Icy\", \"C\": \"Dry\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000983", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/297.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Slippery\", \"B\": \"Icy\", \"C\": \"Dry\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000983", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Airport roadway\", \"C\": \"Service station\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000984", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Airport roadway\", \"C\": \"Service station\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000984", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Warm haze\", \"B\": \"Blinding sunlight\", \"C\": \"Clear\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000985", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Warm haze\", \"B\": \"Blinding sunlight\", \"C\": \"Clear\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000985", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000986", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000986", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Stormy\", \"C\": \"Strong wind\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000987", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Stormy\", \"C\": \"Strong wind\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000987", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Flashing red\", \"C\": \"Not visible\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000988", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Flashing red\", \"C\": \"Not visible\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000988", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Muddy\", \"C\": \"Blocked by obstacle\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000989", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/305.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Muddy\", \"C\": \"Blocked by obstacle\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000989", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Urban\", \"C\": \"Market street\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000990", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Urban\", \"C\": \"Market street\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000990", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in the scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Strong shadows\", \"C\": \"Hazy\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000991", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in the scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Strong shadows\", \"C\": \"Hazy\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000991", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in the scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Accident blockage\", \"C\": \"Lane merging\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000992", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in the scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Accident blockage\", \"C\": \"Lane merging\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000992", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Windy\", \"C\": \"Haze\", \"D\": \"Dusty\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000993", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Windy\", \"C\": \"Haze\", \"D\": \"Dusty\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000993", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"None\", \"C\": \"Not visible\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000994", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"None\", \"C\": \"Not visible\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000994", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Dry\", \"C\": \"Cracked asphalt\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000995", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/314.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Dry\", \"C\": \"Cracked asphalt\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000995", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000996", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000996", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the scene illuminated? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Partly cloudy\", \"C\": \"Dawn\", \"D\": \"Glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000997", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the scene illuminated? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Partly cloudy\", \"C\": \"Dawn\", \"D\": \"Glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000997", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Heavy\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000998", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Heavy\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000998", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Stormy\", \"C\": \"Snowstorm\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000999", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Stormy\", \"C\": \"Snowstorm\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:000999", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Covered by tree\", \"C\": \"Countdown timer visible\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001000", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Covered by tree\", \"C\": \"Countdown timer visible\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001000", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rough surface\", \"C\": \"Cracked asphalt\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001001", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/323.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rough surface\", \"C\": \"Cracked asphalt\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001001", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Parking lot\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001002", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Parking lot\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001002", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the scene illuminated? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Under tunnel\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001003", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the scene illuminated? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Under tunnel\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001003", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001004", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001004", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Sunny\", \"C\": \"Windy\", \"D\": \"Snowstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001005", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Sunny\", \"C\": \"Windy\", \"D\": \"Snowstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001005", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Reflected signal\", \"C\": \"Flashing amber\", \"D\": \"Hidden\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001006", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Reflected signal\", \"C\": \"Flashing amber\", \"D\": \"Hidden\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001006", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Construction debris\", \"C\": \"Icy\", \"D\": \"Potholes\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001007", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/332.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Construction debris\", \"C\": \"Icy\", \"D\": \"Potholes\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001007", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Forest road\", \"C\": \"Airport roadway\", \"D\": \"Avenue\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001008", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Forest road\", \"C\": \"Airport roadway\", \"D\": \"Avenue\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Avenue" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001008", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Golden\", \"B\": \"Flashing lights\", \"C\": \"Overexposed\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001009", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Golden\", \"B\": \"Flashing lights\", \"C\": \"Overexposed\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Golden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001009", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001010", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001010", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Sleet\", \"C\": \"Dry\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001011", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Sleet\", \"C\": \"Dry\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001011", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Reflected signal\", \"C\": \"Unknown\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001012", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Reflected signal\", \"C\": \"Unknown\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001012", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Smooth\", \"C\": \"Sandy\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001013", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/341.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Smooth\", \"C\": \"Sandy\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001013", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001014", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001014", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Under tunnel\", \"C\": \"Dim streetlight visibility\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001015", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Under tunnel\", \"C\": \"Dim streetlight visibility\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001015", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001016", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001016", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Hazy\", \"C\": \"Rainy\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001017", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Hazy\", \"C\": \"Rainy\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001017", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Covered by tree\", \"C\": \"Not visible\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001018", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Covered by tree\", \"C\": \"Not visible\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001018", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Icy\", \"C\": \"Oil slick\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001019", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/350.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Icy\", \"C\": \"Oil slick\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001019", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Airport roadway\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001020", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Airport roadway\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001020", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Clear\", \"C\": \"Hazy\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001021", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Clear\", \"C\": \"Hazy\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001021", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001022", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001022", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Haze\", \"C\": \"Smoggy\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001023", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Haze\", \"C\": \"Smoggy\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001023", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the image? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Blinking green\", \"C\": \"Green\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001024", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the image? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Blinking green\", \"C\": \"Green\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001024", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Flooded\", \"C\": \"Dry\", \"D\": \"Leaf-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001025", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/36.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Flooded\", \"C\": \"Dry\", \"D\": \"Leaf-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001025", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Coastal road\", \"C\": \"Harbor area\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001026", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Coastal road\", \"C\": \"Harbor area\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001026", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Clear\", \"C\": \"Strong shadows\", \"D\": \"Under tunnel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001027", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Clear\", \"C\": \"Strong shadows\", \"D\": \"Under tunnel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001027", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Lane merging\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001028", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Lane merging\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001028", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Dry\", \"C\": \"Thunderstorm\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001029", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Dry\", \"C\": \"Thunderstorm\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001029", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Hidden\", \"B\": \"Blinking green\", \"C\": \"Yellow\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001030", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Hidden\", \"B\": \"Blinking green\", \"C\": \"Yellow\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001030", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Slippery\", \"C\": \"Dry\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001031", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/369.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Slippery\", \"C\": \"Dry\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001031", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Rural\", \"C\": \"Airport roadway\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001032", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Rural\", \"C\": \"Airport roadway\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001032", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Overexposed\", \"C\": \"Hazy\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001033", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Overexposed\", \"C\": \"Hazy\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001033", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001034", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001034", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Snowstorm\", \"C\": \"Smoggy\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001035", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Snowstorm\", \"C\": \"Smoggy\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001035", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Not visible\", \"C\": \"Covered by tree\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001036", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Not visible\", \"C\": \"Covered by tree\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001036", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Dry\", \"C\": \"Uneven\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001037", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/378.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Dry\", \"C\": \"Uneven\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001037", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Urban\", \"C\": \"Airport roadway\", \"D\": \"Campus area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001038", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Urban\", \"C\": \"Airport roadway\", \"D\": \"Campus area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001038", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dawn\", \"C\": \"Dusk\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001039", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dawn\", \"C\": \"Dusk\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001039", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001040", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001040", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Dry\", \"C\": \"Smoggy\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001041", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Dry\", \"C\": \"Smoggy\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001041", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Flashing red\", \"C\": \"Dim / unclear\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001042", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Flashing red\", \"C\": \"Dim / unclear\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001042", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Rough surface\", \"C\": \"Dry\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001043", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/387.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Rough surface\", \"C\": \"Dry\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001043", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001044", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001044", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Foggy\", \"C\": \"Under tunnel\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001045", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Foggy\", \"C\": \"Under tunnel\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001045", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Rush hour\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001046", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like ahead? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Rush hour\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001046", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Strong wind\", \"C\": \"Flurries\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001047", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Strong wind\", \"C\": \"Flurries\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001047", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing amber\", \"C\": \"Flashing yellow\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001048", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing amber\", \"C\": \"Flashing yellow\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001048", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Flooded\", \"C\": \"Potholes\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001049", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/396.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Flooded\", \"C\": \"Potholes\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001049", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Coastal road\", \"D\": \"Campus area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001050", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Coastal road\", \"D\": \"Campus area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001050", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Blinding sunlight\", \"C\": \"Warm haze\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001051", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Blinding sunlight\", \"C\": \"Warm haze\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001051", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001052", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001052", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Clear\", \"C\": \"Flurries\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001053", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Clear\", \"C\": \"Flurries\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001053", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Flashing amber\", \"C\": \"Unknown\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001054", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Flashing amber\", \"C\": \"Unknown\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001054", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Newly paved\", \"C\": \"Dry\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001055", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/404.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Newly paved\", \"C\": \"Dry\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001055", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Construction zone\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001056", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Construction zone\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001056", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Overexposed\", \"C\": \"Dusk\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001057", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Overexposed\", \"C\": \"Dusk\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001057", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition ahead? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Congested\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001058", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition ahead? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Congested\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001058", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Sunny\", \"C\": \"Cloudy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001059", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Sunny\", \"C\": \"Cloudy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001059", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Not visible\", \"C\": \"Off / malfunctioning\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001060", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Not visible\", \"C\": \"Off / malfunctioning\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001060", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Smooth\", \"C\": \"Wet\", \"D\": \"Potholes\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001061", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/413.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Smooth\", \"C\": \"Wet\", \"D\": \"Potholes\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001061", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Urban\", \"C\": \"Bridge\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001062", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Urban\", \"C\": \"Bridge\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001062", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Backlit\", \"C\": \"Under tunnel\", \"D\": \"Reflected glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001063", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Backlit\", \"C\": \"Under tunnel\", \"D\": \"Reflected glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001063", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001064", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001064", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Dry\", \"C\": \"Thunderstorm\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001065", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Dry\", \"C\": \"Thunderstorm\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001065", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Yellow\", \"C\": \"Countdown timer visible\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001066", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Yellow\", \"C\": \"Countdown timer visible\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001066", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Sandy\", \"C\": \"Dry\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001067", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/422.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Sandy\", \"C\": \"Dry\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001067", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Subway entrance\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001068", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Subway entrance\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001068", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the image in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dim streetlight visibility\", \"C\": \"Dawn\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001069", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the image in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dim streetlight visibility\", \"C\": \"Dawn\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001069", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001070", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001070", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Cloudy\", \"C\": \"Stormy\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001071", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Cloudy\", \"C\": \"Stormy\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001071", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Red\", \"C\": \"Flashing amber\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001072", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Red\", \"C\": \"Flashing amber\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001072", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001073", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/431.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001073", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Campus area\", \"D\": \"Construction zone\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001074", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Campus area\", \"D\": \"Construction zone\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001074", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Clear\", \"C\": \"Strong shadows\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001075", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Clear\", \"C\": \"Strong shadows\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001075", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Heavy\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001076", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Heavy\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001076", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Strong wind\", \"C\": \"Cloudy\", \"D\": \"Dusty\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001077", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Strong wind\", \"C\": \"Cloudy\", \"D\": \"Dusty\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Cloudy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001077", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"None\", \"C\": \"Countdown timer visible\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001078", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"None\", \"C\": \"Countdown timer visible\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001078", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Cracked asphalt\", \"C\": \"Partially damaged\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001079", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/440.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Cracked asphalt\", \"C\": \"Partially damaged\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001079", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Bridge\", \"C\": \"Construction zone\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001080", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Bridge\", \"C\": \"Construction zone\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001080", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Clear\", \"C\": \"Dawn\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001081", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Clear\", \"C\": \"Dawn\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001081", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001082", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001082", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Flurries\", \"C\": \"Clear\", \"D\": \"Thunderstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001083", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Flurries\", \"C\": \"Clear\", \"D\": \"Thunderstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001083", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Green\", \"C\": \"Not visible\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001084", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Green\", \"C\": \"Not visible\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001084", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Cracked asphalt\", \"C\": \"Bumpy\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001085", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/45.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Cracked asphalt\", \"C\": \"Bumpy\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001085", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Parking lot\", \"C\": \"Suburban\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001086", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Parking lot\", \"C\": \"Suburban\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001086", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Hazy\", \"C\": \"Strong shadows\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001087", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Hazy\", \"C\": \"Strong shadows\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001087", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Detour traffic\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001088", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Detour traffic\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001088", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Dusty\", \"C\": \"Dry\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001089", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Dusty\", \"C\": \"Dry\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001089", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Not visible\", \"C\": \"Green\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001090", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Not visible\", \"C\": \"Green\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001090", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Slippery\", \"C\": \"Bumpy\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001091", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/459.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Slippery\", \"C\": \"Bumpy\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001091", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Shopping district\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001092", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Shopping district\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001092", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Dim streetlight visibility\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001093", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Dim streetlight visibility\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001093", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001094", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001094", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Foggy\", \"C\": \"Gloomy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001095", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Foggy\", \"C\": \"Gloomy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001095", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Absent\", \"C\": \"Yellow\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001096", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Absent\", \"C\": \"Yellow\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001096", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Construction debris\", \"C\": \"Wet\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001097", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/468.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Construction debris\", \"C\": \"Wet\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001097", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Urban\", \"C\": \"Forest road\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001098", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Urban\", \"C\": \"Forest road\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001098", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hazy\", \"C\": \"Overexposed\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001099", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hazy\", \"C\": \"Overexposed\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001099", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001100", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001100", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Sunny\", \"C\": \"Hailing\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001101", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Sunny\", \"C\": \"Hailing\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001101", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Yellow\", \"C\": \"Reflected signal\", \"D\": \"No\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001102", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Yellow\", \"C\": \"Reflected signal\", \"D\": \"No\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001102", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Uneven\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001103", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/477.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Uneven\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001103", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Market street\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001104", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Market street\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001104", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Blurry\", \"B\": \"Under tunnel\", \"C\": \"Glare\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001105", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Blurry\", \"B\": \"Under tunnel\", \"C\": \"Glare\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Blurry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001105", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001106", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001106", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Sleet\", \"C\": \"Hazy\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001107", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Sleet\", \"C\": \"Hazy\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001107", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Not visible\", \"C\": \"Temporary road sign light\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001108", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Not visible\", \"C\": \"Temporary road sign light\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001108", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Dry\", \"C\": \"Bumpy\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001109", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/486.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Dry\", \"C\": \"Bumpy\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001109", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Parking lot\", \"C\": \"Shopping district\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001110", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Parking lot\", \"C\": \"Shopping district\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001110", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Under tunnel\", \"C\": \"Dim streetlight visibility\", \"D\": \"Blurry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001111", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Under tunnel\", \"C\": \"Dim streetlight visibility\", \"D\": \"Blurry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Blurry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001111", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001112", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001112", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Dusty\", \"C\": \"Strong wind\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001113", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Dusty\", \"C\": \"Strong wind\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001113", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Yellow\", \"C\": \"Flashing yellow\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001114", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Yellow\", \"C\": \"Flashing yellow\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001114", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Sandy\", \"C\": \"Snow-covered\", \"D\": \"Gravel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001115", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/495.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Sandy\", \"C\": \"Snow-covered\", \"D\": \"Gravel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001115", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001116", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001116", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Flashing lights\", \"C\": \"Low\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001117", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Flashing lights\", \"C\": \"Low\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001117", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001118", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001118", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Smoggy\", \"C\": \"Mist\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001119", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Smoggy\", \"C\": \"Mist\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001119", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing amber\", \"C\": \"Pedestrian signal\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001120", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing amber\", \"C\": \"Pedestrian signal\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001120", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Dry\", \"C\": \"Oil slick\", \"D\": \"Icy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001121", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/503.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Dry\", \"C\": \"Oil slick\", \"D\": \"Icy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001121", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Coastal road\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001122", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Coastal road\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001122", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity of the scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Glare\", \"C\": \"Clear\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001123", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity of the scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Glare\", \"C\": \"Clear\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001123", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001124", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001124", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Thunderstorm\", \"C\": \"Windy\", \"D\": \"Overcast\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001125", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Thunderstorm\", \"C\": \"Windy\", \"D\": \"Overcast\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001125", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Pedestrian signal\", \"C\": \"None\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001126", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Pedestrian signal\", \"C\": \"None\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001126", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Bumpy\", \"C\": \"Snow-covered\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001127", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/512.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Bumpy\", \"C\": \"Snow-covered\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001127", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Coastal road\", \"C\": \"Construction zone\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001128", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Coastal road\", \"C\": \"Construction zone\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001128", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Hazy\", \"C\": \"Strong shadows\", \"D\": \"Under tunnel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001129", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Hazy\", \"C\": \"Strong shadows\", \"D\": \"Under tunnel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001129", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001130", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001130", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Windy\", \"C\": \"Hazy\", \"D\": \"Strong wind\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001131", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Windy\", \"C\": \"Hazy\", \"D\": \"Strong wind\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001131", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"No\", \"C\": \"Green\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001132", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"No\", \"C\": \"Green\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001132", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Wet\", \"C\": \"Slippery\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001133", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/521.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Wet\", \"C\": \"Slippery\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001133", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Urban\", \"C\": \"Service station\", \"D\": \"Construction zone\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001134", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Urban\", \"C\": \"Service station\", \"D\": \"Construction zone\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001134", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Warm\", \"C\": \"Rain-blurred\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001135", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Warm\", \"C\": \"Rain-blurred\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001135", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle activity? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001136", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle activity? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001136", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Snowy\", \"C\": \"Rainy\", \"D\": \"Overcast\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001137", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Snowy\", \"C\": \"Rainy\", \"D\": \"Overcast\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Overcast" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001137", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light active or visible? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Blinking green\", \"C\": \"Flashing yellow\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001138", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light active or visible? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Blinking green\", \"C\": \"Flashing yellow\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001138", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snow-covered\", \"C\": \"Oil slick\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001139", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/530.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snow-covered\", \"C\": \"Oil slick\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001139", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Campus area\", \"C\": \"Construction zone\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001140", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Campus area\", \"C\": \"Construction zone\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001140", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Dawn\", \"C\": \"Backlit\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001141", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Dawn\", \"C\": \"Backlit\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001141", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001142", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001142", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Mist\", \"C\": \"Flurries\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001143", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Mist\", \"C\": \"Flurries\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001143", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"None\", \"C\": \"Not visible\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001144", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"None\", \"C\": \"Not visible\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001144", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Debris scattered\", \"C\": \"Cracked asphalt\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001145", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/54.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Debris scattered\", \"C\": \"Cracked asphalt\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001145", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Suburban\", \"C\": \"Rural\", \"D\": \"Residential\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001146", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Suburban\", \"C\": \"Rural\", \"D\": \"Residential\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001146", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Dim streetlight visibility\", \"C\": \"Partly cloudy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001147", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Dim streetlight visibility\", \"C\": \"Partly cloudy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001147", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001148", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001148", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Cloudy\", \"C\": \"Sleet\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001149", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Cloudy\", \"C\": \"Sleet\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001149", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Green\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001150", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Green\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001150", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Dry\", \"C\": \"Cracked asphalt\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001151", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/549.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Dry\", \"C\": \"Cracked asphalt\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001151", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Campus area\", \"C\": \"Coastal road\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001152", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Campus area\", \"C\": \"Coastal road\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001152", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Smoky\", \"C\": \"Reflected glare\", \"D\": \"Low\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001153", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Smoky\", \"C\": \"Reflected glare\", \"D\": \"Low\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001153", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001154", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001154", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Dry\", \"C\": \"Sandstorm\", \"D\": \"Thunderstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001155", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Dry\", \"C\": \"Sandstorm\", \"D\": \"Thunderstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001155", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Red\", \"C\": \"None\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001156", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Red\", \"C\": \"None\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001156", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Paved\", \"B\": \"Bumpy\", \"C\": \"Partially damaged\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001157", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/558.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Paved\", \"B\": \"Bumpy\", \"C\": \"Partially damaged\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Paved" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001157", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Parking lot\", \"C\": \"Shopping district\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001158", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Parking lot\", \"C\": \"Shopping district\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001158", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Hazy\", \"C\": \"Clear\", \"D\": \"Glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001159", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Hazy\", \"C\": \"Clear\", \"D\": \"Glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001159", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there heavy traffic on this road? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001160", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there heavy traffic on this road? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001160", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Smoggy\", \"C\": \"Drizzling\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001161", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Smoggy\", \"C\": \"Drizzling\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001161", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Temporary road sign light\", \"C\": \"Unknown\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001162", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Temporary road sign light\", \"C\": \"Unknown\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001162", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Newly paved\", \"C\": \"Construction debris\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001163", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/567.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Newly paved\", \"C\": \"Construction debris\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001163", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Construction zone\", \"C\": \"Urban\", \"D\": \"Parking lot\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001164", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Construction zone\", \"C\": \"Urban\", \"D\": \"Parking lot\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001164", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Rain-blurred\", \"C\": \"Partly cloudy\", \"D\": \"Warm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001165", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Rain-blurred\", \"C\": \"Partly cloudy\", \"D\": \"Warm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001165", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001166", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001166", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Hailing\", \"C\": \"Dusty\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001167", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Hailing\", \"C\": \"Dusty\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001167", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Red\", \"C\": \"Not visible\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001168", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Red\", \"C\": \"Not visible\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001168", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Oil slick\", \"C\": \"Snow-covered\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001169", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/576.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Oil slick\", \"C\": \"Snow-covered\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001169", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Roundabout\", \"C\": \"Airport roadway\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001170", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Roundabout\", \"C\": \"Airport roadway\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001170", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Clear\", \"C\": \"Strong shadows\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001171", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Clear\", \"C\": \"Strong shadows\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001171", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001172", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001172", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Snowy\", \"C\": \"Dry\", \"D\": \"Dusty\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001173", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Snowy\", \"C\": \"Dry\", \"D\": \"Dusty\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001173", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing yellow\", \"C\": \"Yellow\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001174", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing yellow\", \"C\": \"Yellow\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001174", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Dry\", \"C\": \"Rough surface\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001175", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/585.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Dry\", \"C\": \"Rough surface\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001175", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Market street\", \"C\": \"Bridge\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001176", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Market street\", \"C\": \"Bridge\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001176", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the scene illuminated? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low light\", \"C\": \"Foggy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001177", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the scene illuminated? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low light\", \"C\": \"Foggy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001177", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Moderate\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001178", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Moderate\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001178", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Foggy\", \"C\": \"Sleet\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001179", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Foggy\", \"C\": \"Sleet\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001179", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Absent\", \"C\": \"Off / malfunctioning\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001180", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Absent\", \"C\": \"Off / malfunctioning\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001180", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Potholes\", \"C\": \"Dry\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001181", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/594.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Potholes\", \"C\": \"Dry\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001181", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Service station\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001182", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Service station\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001182", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Backlit\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001183", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Backlit\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001183", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001184", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001184", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Dry\", \"C\": \"Gloomy\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001185", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Dry\", \"C\": \"Gloomy\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001185", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Temporary road sign light\", \"C\": \"Reflected signal\", \"D\": \"Hidden\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001186", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Temporary road sign light\", \"C\": \"Reflected signal\", \"D\": \"Hidden\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001186", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Smooth\", \"C\": \"Snow-covered\", \"D\": \"Cracked asphalt\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001187", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/602.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Smooth\", \"C\": \"Snow-covered\", \"D\": \"Cracked asphalt\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001187", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Shopping district\", \"C\": \"Urban\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001188", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Shopping district\", \"C\": \"Urban\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001188", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility affected in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Low light\", \"C\": \"Limited\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001189", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility affected in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Low light\", \"C\": \"Limited\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Limited" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001189", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Accident blockage\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001190", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Accident blockage\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001190", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Hazy\", \"C\": \"Sleet\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001191", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Hazy\", \"C\": \"Sleet\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001191", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Flashing red\", \"C\": \"Covered by tree\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001192", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Flashing red\", \"C\": \"Covered by tree\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001192", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Potholes\", \"C\": \"Smooth\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001193", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/611.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Potholes\", \"C\": \"Smooth\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001193", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Parking lot\", \"C\": \"Construction zone\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001194", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Parking lot\", \"C\": \"Construction zone\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001194", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hazy\", \"C\": \"Dusk\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001195", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hazy\", \"C\": \"Dusk\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001195", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Detour traffic\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001196", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Detour traffic\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001196", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Cloudy\", \"C\": \"Windy\", \"D\": \"Sleet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001197", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Cloudy\", \"C\": \"Windy\", \"D\": \"Sleet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Cloudy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001197", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Blinking green\", \"C\": \"Not visible\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001198", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Blinking green\", \"C\": \"Not visible\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001198", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Partially damaged\", \"C\": \"Rough surface\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001199", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/620.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Partially damaged\", \"C\": \"Rough surface\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001199", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Urban\", \"C\": \"Highway\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001200", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Urban\", \"C\": \"Highway\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001200", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Backlit\", \"C\": \"Foggy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001201", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Backlit\", \"C\": \"Foggy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001201", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Moderate\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001202", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Moderate\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001202", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Clear\", \"C\": \"Snowstorm\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001203", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Clear\", \"C\": \"Snowstorm\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001203", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Off / malfunctioning\", \"C\": \"Not visible\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001204", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Off / malfunctioning\", \"C\": \"Not visible\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001204", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Rough surface\", \"C\": \"Wet\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001205", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/63.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Rough surface\", \"C\": \"Wet\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001205", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Service station\", \"D\": \"Construction zone\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001206", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Service station\", \"D\": \"Construction zone\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001206", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Reflected glare\", \"C\": \"Excellent\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001207", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Reflected glare\", \"C\": \"Excellent\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Excellent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001207", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001208", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001208", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Sleet\", \"C\": \"Foggy\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001209", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Sleet\", \"C\": \"Foggy\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001209", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Hidden\", \"C\": \"Red\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001210", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Hidden\", \"C\": \"Red\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001210", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001211", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/639.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Dry\", \"C\": \"Wet\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001211", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Forest road\", \"C\": \"Parking lot\", \"D\": \"City street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001212", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Forest road\", \"C\": \"Parking lot\", \"D\": \"City street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "City street" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001212", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Soft light\", \"B\": \"Reflected glare\", \"C\": \"Strong shadows\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001213", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Soft light\", \"B\": \"Reflected glare\", \"C\": \"Strong shadows\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Soft light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001213", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001214", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001214", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any weather conditions affecting the scene? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Snowy\", \"C\": \"Strong wind\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001215", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any weather conditions affecting the scene? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Snowy\", \"C\": \"Strong wind\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001215", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Flashing amber\", \"C\": \"None\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001216", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Flashing amber\", \"C\": \"None\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001216", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Debris scattered\", \"C\": \"Partially damaged\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001217", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/648.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Debris scattered\", \"C\": \"Partially damaged\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001217", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Highway\", \"C\": \"Service station\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001218", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Highway\", \"C\": \"Service station\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001218", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Low\", \"C\": \"Overexposed\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001219", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Low\", \"C\": \"Overexposed\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001219", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Congested\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001220", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Congested\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001220", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Stormy\", \"C\": \"Clear\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001221", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Stormy\", \"C\": \"Clear\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001221", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing red\", \"C\": \"Red\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001222", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing red\", \"C\": \"Red\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001222", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001223", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/657.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001223", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Parking lot\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001224", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Parking lot\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001224", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Backlit\", \"C\": \"Clear\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001225", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Backlit\", \"C\": \"Clear\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001225", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001226", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001226", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Cloudy\", \"C\": \"Dusty\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001227", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Cloudy\", \"C\": \"Dusty\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Overcast" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001227", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Temporary road sign light\", \"C\": \"Dim / unclear\", \"D\": \"Absent\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001228", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Temporary road sign light\", \"C\": \"Dim / unclear\", \"D\": \"Absent\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001228", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Partially damaged\", \"C\": \"Sandy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001229", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/666.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Partially damaged\", \"C\": \"Sandy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001229", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Rural\", \"C\": \"Campus area\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001230", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Rural\", \"C\": \"Campus area\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001230", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Blinding sunlight\", \"C\": \"Dusk\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001231", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Blinding sunlight\", \"C\": \"Dusk\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001231", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001232", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001232", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Stormy\", \"C\": \"Sunny\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001233", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Stormy\", \"C\": \"Sunny\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001233", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Covered by tree\", \"C\": \"Unknown\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001234", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Covered by tree\", \"C\": \"Unknown\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001234", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Blocked by obstacle\", \"C\": \"Gravel\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001235", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/675.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Debris scattered\", \"B\": \"Blocked by obstacle\", \"C\": \"Gravel\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001235", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Forest road\", \"C\": \"Suburban\", \"D\": \"Residential\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001236", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Forest road\", \"C\": \"Suburban\", \"D\": \"Residential\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001236", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity of the scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Rain-blurred\", \"C\": \"Overexposed\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001237", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity of the scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Rain-blurred\", \"C\": \"Overexposed\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001237", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001238", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001238", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Thunderstorm\", \"C\": \"Clear\", \"D\": \"Sleet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001239", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Thunderstorm\", \"C\": \"Clear\", \"D\": \"Sleet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001239", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the image? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing amber\", \"C\": \"Green\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001240", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the image? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Flashing amber\", \"C\": \"Green\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001240", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Newly paved\", \"C\": \"Partially damaged\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001241", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/684.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Newly paved\", \"C\": \"Partially damaged\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001241", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Suburban\", \"C\": \"Market street\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001242", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Suburban\", \"C\": \"Market street\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001242", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Clear\", \"C\": \"Hazy\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001243", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Clear\", \"C\": \"Hazy\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001243", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on the road? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Lane merging\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001244", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on the road? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Lane merging\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001244", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rainy\", \"C\": \"Cloudy\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001245", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rainy\", \"C\": \"Cloudy\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001245", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Absent\", \"C\": \"Dim / unclear\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001246", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Absent\", \"C\": \"Dim / unclear\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001246", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Icy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001247", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/693.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Icy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001247", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Airport roadway\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001248", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Airport roadway\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001248", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Flashing lights\", \"C\": \"Hazy\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001249", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Flashing lights\", \"C\": \"Hazy\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001249", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001250", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001250", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Hazy\", \"C\": \"Snowy\", \"D\": \"Strong wind\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001251", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Hazy\", \"C\": \"Snowy\", \"D\": \"Strong wind\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001251", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001252", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001252", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Snowy\", \"C\": \"Rough surface\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001253", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/701.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Snowy\", \"C\": \"Rough surface\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001253", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Parking lot\", \"C\": \"Airport roadway\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001254", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Parking lot\", \"C\": \"Airport roadway\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001254", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Smoky\", \"C\": \"Blinding sunlight\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001255", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Smoky\", \"C\": \"Blinding sunlight\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001255", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition in the scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001256", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition in the scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001256", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Snowstorm\", \"C\": \"Gloomy\", \"D\": \"Sleet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001257", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Snowstorm\", \"C\": \"Gloomy\", \"D\": \"Sleet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001257", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Absent\", \"C\": \"Off / malfunctioning\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001258", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Absent\", \"C\": \"Off / malfunctioning\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001258", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001259", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/710.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001259", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Airport roadway\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001260", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Airport roadway\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001260", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Blinding sunlight\", \"C\": \"Dawn\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001261", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Blinding sunlight\", \"C\": \"Dawn\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001261", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Heavy\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001262", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Heavy\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001262", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Sunny\", \"C\": \"Snowstorm\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001263", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Sunny\", \"C\": \"Snowstorm\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001263", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this frame? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Not visible\", \"C\": \"Blinking green\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001264", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this frame? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Not visible\", \"C\": \"Blinking green\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001264", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described in the scene? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001265", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/72.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described in the scene? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001265", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Coastal road\", \"C\": \"Service station\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001266", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Coastal road\", \"C\": \"Service station\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001266", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Smoky\", \"C\": \"Clear\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001267", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Smoky\", \"C\": \"Clear\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001267", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition on the road? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001268", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition on the road? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001268", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Sandstorm\", \"C\": \"Strong wind\", \"D\": \"Sleet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001269", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Sandstorm\", \"C\": \"Strong wind\", \"D\": \"Sleet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001269", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Blinking green\", \"C\": \"Absent\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001270", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Blinking green\", \"C\": \"Absent\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001270", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Dry\", \"C\": \"Gravel\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001271", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/729.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Dry\", \"C\": \"Gravel\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001271", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001272", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001272", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Backlit\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001273", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Backlit\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001273", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Heavy\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001274", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Heavy\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001274", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Haze\", \"C\": \"Snowy\", \"D\": \"Mist\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001275", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Haze\", \"C\": \"Snowy\", \"D\": \"Mist\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001275", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Reflected signal\", \"C\": \"Green\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001276", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Reflected signal\", \"C\": \"Green\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001276", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Slippery\", \"B\": \"Gravel\", \"C\": \"Uneven\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001277", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/738.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Slippery\", \"B\": \"Gravel\", \"C\": \"Uneven\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001277", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001278", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001278", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Flashing lights\", \"C\": \"Low light\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001279", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Flashing lights\", \"C\": \"Low light\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001279", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001280", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001280", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snowstorm\", \"C\": \"Hailing\", \"D\": \"Mist\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001281", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snowstorm\", \"C\": \"Hailing\", \"D\": \"Mist\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001281", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Flashing yellow\", \"C\": \"No\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001282", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Flashing yellow\", \"C\": \"No\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001282", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Dry\", \"C\": \"Slippery\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001283", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/747.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Dry\", \"C\": \"Slippery\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001283", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Subway entrance\", \"C\": \"Rural\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001284", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Subway entrance\", \"C\": \"Rural\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001284", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Partly cloudy\", \"C\": \"Glare\", \"D\": \"Dim\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001285", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Partly cloudy\", \"C\": \"Glare\", \"D\": \"Dim\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dim" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001285", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001286", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001286", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Clear\", \"C\": \"Overcast\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001287", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Clear\", \"C\": \"Overcast\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001287", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Yellow\", \"C\": \"Flashing yellow\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001288", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Yellow\", \"C\": \"Flashing yellow\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001288", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Wet\", \"C\": \"Icy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001289", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/756.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Wet\", \"C\": \"Icy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001289", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Construction zone\", \"C\": \"Coastal road\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001290", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Construction zone\", \"C\": \"Coastal road\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001290", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene described? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Hazy\", \"C\": \"Reflected glare\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001291", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene described? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Hazy\", \"C\": \"Reflected glare\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001291", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any visible vehicle movement in the scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Moderate\", \"C\": \"None\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001292", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any visible vehicle movement in the scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Moderate\", \"C\": \"None\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001292", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is implied by the scene? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Rainy\", \"C\": \"Snowstorm\", \"D\": \"Haze\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001293", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is implied by the scene? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Rainy\", \"C\": \"Snowstorm\", \"D\": \"Haze\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001293", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"No\", \"C\": \"Red\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001294", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"No\", \"C\": \"Red\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001294", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Uneven\", \"C\": \"Dry\", \"D\": \"Debris scattered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001295", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/765.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Uneven\", \"C\": \"Dry\", \"D\": \"Debris scattered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001295", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Bridge\", \"C\": \"Suburban\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001296", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Bridge\", \"C\": \"Suburban\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001296", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Warm light\", \"B\": \"Partly cloudy\", \"C\": \"Dim streetlight visibility\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001297", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Warm light\", \"B\": \"Partly cloudy\", \"C\": \"Dim streetlight visibility\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001297", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Congested\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001298", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Congested\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001298", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Dry\", \"C\": \"Stormy\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001299", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Dry\", \"C\": \"Stormy\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001299", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Covered by tree\", \"C\": \"Blinking green\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001300", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Covered by tree\", \"C\": \"Blinking green\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001300", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Construction debris\", \"C\": \"Wet\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001301", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/774.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Construction debris\", \"C\": \"Wet\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001301", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Market street\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001302", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Market street\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001302", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Overexposed\", \"C\": \"Clear\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001303", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Overexposed\", \"C\": \"Clear\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001303", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there traffic on the road? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Congested\", \"C\": \"None\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001304", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there traffic on the road? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Congested\", \"C\": \"None\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001304", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Windy\", \"C\": \"Flurries\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001305", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Windy\", \"C\": \"Flurries\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001305", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Yellow\", \"C\": \"Reflected signal\", \"D\": \"No\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001306", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Yellow\", \"C\": \"Reflected signal\", \"D\": \"No\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001306", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Smooth\", \"C\": \"Oil slick\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001307", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/783.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Smooth\", \"C\": \"Oil slick\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001307", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Harbor area\", \"C\": \"Suburban\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001308", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Harbor area\", \"C\": \"Suburban\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001308", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in the scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Clear\", \"C\": \"Reflected glare\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001309", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in the scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Clear\", \"C\": \"Reflected glare\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001309", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001310", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001310", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Dry\", \"C\": \"Snowstorm\", \"D\": \"Overcast\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001311", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Dry\", \"C\": \"Snowstorm\", \"D\": \"Overcast\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001311", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"None\", \"C\": \"Off / malfunctioning\", \"D\": \"No\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001312", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"None\", \"C\": \"Off / malfunctioning\", \"D\": \"No\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001312", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Smooth\", \"C\": \"Partially damaged\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001313", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/792.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Smooth\", \"C\": \"Partially damaged\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001313", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of street is this? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001314", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of street is this? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001314", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Excellent\", \"C\": \"Rain-blurred\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001315", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Excellent\", \"C\": \"Rain-blurred\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Excellent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001315", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001316", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001316", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Gloomy\", \"C\": \"Snowy\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001317", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Gloomy\", \"C\": \"Snowy\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001317", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Flashing red\", \"C\": \"Absent\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001318", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Flashing red\", \"C\": \"Absent\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001318", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Debris scattered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001319", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/800.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Debris scattered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001319", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Service station\", \"C\": \"Coastal road\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001320", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Service station\", \"C\": \"Coastal road\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001320", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Reflected glare\", \"C\": \"Clear\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001321", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Reflected glare\", \"C\": \"Clear\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001321", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Lane merging\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001322", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Lane merging\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001322", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Dusty\", \"C\": \"Haze\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001323", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Dusty\", \"C\": \"Haze\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001323", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing yellow\", \"C\": \"Temporary road sign light\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001324", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing yellow\", \"C\": \"Temporary road sign light\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001324", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Slippery\", \"C\": \"Sandy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001325", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/81.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Slippery\", \"C\": \"Sandy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001325", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Shopping district\", \"C\": \"Harbor area\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001326", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Shopping district\", \"C\": \"Harbor area\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001326", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Rain-blurred\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001327", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Rain-blurred\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001327", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001328", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001328", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Mist\", \"C\": \"Smog\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001329", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Mist\", \"C\": \"Smog\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smog" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001329", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"None\", \"B\": \"No\", \"C\": \"Flashing yellow\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001330", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"None\", \"B\": \"No\", \"C\": \"Flashing yellow\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001330", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rough surface\", \"C\": \"Gravel\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001331", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/819.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rough surface\", \"C\": \"Gravel\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001331", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001332", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001332", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Backlit\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001333", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Backlit\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001333", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition around the road? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001334", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition around the road? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001334", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Dry\", \"C\": \"Overcast\", \"D\": \"Drizzling\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001335", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Dry\", \"C\": \"Overcast\", \"D\": \"Drizzling\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001335", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001336", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001336", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Smooth\", \"C\": \"Rough surface\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001337", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/828.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Smooth\", \"C\": \"Rough surface\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001337", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Suburban\", \"C\": \"Market street\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001338", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Suburban\", \"C\": \"Market street\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001338", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Warm light\", \"B\": \"Smoky\", \"C\": \"Blinding sunlight\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001339", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Warm light\", \"B\": \"Smoky\", \"C\": \"Blinding sunlight\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001339", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001340", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001340", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Smoggy\", \"C\": \"Drizzling\", \"D\": \"Dusty\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001341", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Smoggy\", \"C\": \"Drizzling\", \"D\": \"Dusty\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001341", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Countdown timer visible\", \"C\": \"Not visible\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001342", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Countdown timer visible\", \"C\": \"Not visible\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001342", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Bumpy\", \"C\": \"Smooth\", \"D\": \"Cracked asphalt\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001343", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/837.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Bumpy\", \"C\": \"Smooth\", \"D\": \"Cracked asphalt\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001343", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Airport roadway\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001344", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Airport roadway\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001344", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Blurred\", \"C\": \"Low light\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001345", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Blurred\", \"C\": \"Low light\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Blurred" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001345", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Lane merging\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001346", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Lane merging\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001346", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Foggy\", \"C\": \"Flurries\", \"D\": \"Mist\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001347", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Foggy\", \"C\": \"Flurries\", \"D\": \"Mist\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001347", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Yellow\", \"C\": \"Countdown timer visible\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001348", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Yellow\", \"C\": \"Countdown timer visible\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001348", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Dry\", \"C\": \"Debris scattered\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001349", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/846.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Dry\", \"C\": \"Debris scattered\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001349", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Urban\", \"C\": \"Rural\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001350", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Urban\", \"C\": \"Rural\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001350", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Under tunnel\", \"C\": \"Hazy\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001351", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Under tunnel\", \"C\": \"Hazy\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001351", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001352", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001352", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is visible in the scene? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Snowstorm\", \"C\": \"Stormy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001353", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is visible in the scene? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Snowstorm\", \"C\": \"Stormy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001353", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in this scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Dim / unclear\", \"C\": \"Yellow\", \"D\": \"Absent\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001354", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in this scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Dim / unclear\", \"C\": \"Yellow\", \"D\": \"Absent\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001354", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Water puddles\", \"C\": \"Oil slick\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001355", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/855.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Water puddles\", \"C\": \"Oil slick\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001355", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Market street\", \"C\": \"Urban\", \"D\": \"Campus area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001356", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Market street\", \"C\": \"Urban\", \"D\": \"Campus area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001356", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dusk\", \"C\": \"Reflected glare\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001357", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dusk\", \"C\": \"Reflected glare\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001357", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Moderate\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001358", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Moderate\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001358", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Dry\", \"C\": \"Foggy\", \"D\": \"Strong wind\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001359", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Dry\", \"C\": \"Foggy\", \"D\": \"Strong wind\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001359", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Temporary road sign light\", \"C\": \"Not visible\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001360", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Temporary road sign light\", \"C\": \"Not visible\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001360", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Smooth\", \"C\": \"Construction debris\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001361", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/864.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Smooth\", \"C\": \"Construction debris\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001361", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Parking lot\", \"C\": \"Rural\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001362", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Parking lot\", \"C\": \"Rural\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001362", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Foggy\", \"C\": \"Smoky\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001363", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Foggy\", \"C\": \"Smoky\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001363", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Heavy\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001364", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Heavy\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001364", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Sandstorm\", \"C\": \"Stormy\", \"D\": \"Haze\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001365", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Sandstorm\", \"C\": \"Stormy\", \"D\": \"Haze\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001365", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Temporary road sign light\", \"C\": \"Covered by tree\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001366", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Temporary road sign light\", \"C\": \"Covered by tree\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001366", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Uneven\", \"C\": \"Snow-covered\", \"D\": \"Debris scattered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001367", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/873.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Uneven\", \"C\": \"Snow-covered\", \"D\": \"Debris scattered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001367", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Campus area\", \"C\": \"Suburban\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001368", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Campus area\", \"C\": \"Suburban\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001368", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Foggy\", \"C\": \"Rain-blurred\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001369", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Foggy\", \"C\": \"Rain-blurred\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001369", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001370", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic movement like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001370", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Clear\", \"C\": \"Dusty\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001371", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Clear\", \"C\": \"Dusty\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001371", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"None\", \"C\": \"Hidden\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001372", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"None\", \"C\": \"Hidden\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001372", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Bumpy\", \"C\": \"Snow-covered\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001373", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/882.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Bumpy\", \"C\": \"Snow-covered\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001373", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Subway entrance\", \"C\": \"Bridge\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001374", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Subway entrance\", \"C\": \"Bridge\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001374", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dawn\", \"C\": \"Reflected glare\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001375", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dawn\", \"C\": \"Reflected glare\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001375", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Detour traffic\", \"C\": \"Lane merging\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001376", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Detour traffic\", \"C\": \"Lane merging\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001376", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Windy\", \"C\": \"Snowy\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001377", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Windy\", \"C\": \"Snowy\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001377", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Unknown\", \"C\": \"Flashing amber\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001378", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Unknown\", \"C\": \"Flashing amber\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001378", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Snowy\", \"C\": \"Slippery\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001379", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/891.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Snowy\", \"C\": \"Slippery\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001379", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Campus area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001380", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Campus area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001380", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Overexposed\", \"C\": \"Glare\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001381", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Overexposed\", \"C\": \"Glare\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001381", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001382", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001382", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Dusty\", \"C\": \"Strong wind\", \"D\": \"Haze\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001383", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Dusty\", \"C\": \"Strong wind\", \"D\": \"Haze\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001383", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Unknown\", \"C\": \"Countdown timer visible\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001384", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Unknown\", \"C\": \"Countdown timer visible\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001384", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Newly paved\", \"C\": \"Snow-covered\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001385", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/90.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Newly paved\", \"C\": \"Snow-covered\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001385", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Urban\", \"C\": \"Bridge\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001386", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Urban\", \"C\": \"Bridge\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001386", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Low light\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001387", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Low light\", \"C\": \"Hazy\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001387", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001388", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001388", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Gloomy\", \"C\": \"Sunny\", \"D\": \"Sleet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001389", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Gloomy\", \"C\": \"Sunny\", \"D\": \"Sleet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001389", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"None\", \"C\": \"Absent\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001390", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"None\", \"C\": \"Absent\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001390", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Flooded\", \"C\": \"Cracked asphalt\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001391", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/909.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Flooded\", \"C\": \"Cracked asphalt\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001391", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Market street\", \"C\": \"Urban\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001392", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Market street\", \"C\": \"Urban\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001392", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Glare\", \"C\": \"Clear\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001393", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Glare\", \"C\": \"Clear\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001393", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001394", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001394", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Hailing\", \"C\": \"Cloudy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001395", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Hailing\", \"C\": \"Cloudy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001395", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light currently active? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Covered by tree\", \"C\": \"Unknown\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001396", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light currently active? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Covered by tree\", \"C\": \"Unknown\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001396", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001397", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/918.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001397", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Shopping district\", \"C\": \"Harbor area\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001398", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Shopping district\", \"C\": \"Harbor area\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001398", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Flashing lights\", \"C\": \"Reflected glare\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001399", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Flashing lights\", \"C\": \"Reflected glare\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001399", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001400", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001400", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Sandstorm\", \"C\": \"Dry\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001401", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Sandstorm\", \"C\": \"Dry\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001401", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"None\", \"C\": \"Dim / unclear\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001402", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"None\", \"C\": \"Dim / unclear\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001402", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Newly paved\", \"C\": \"Oil slick\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001403", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/927.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Newly paved\", \"C\": \"Oil slick\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001403", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001404", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001404", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the visual environment? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low light\", \"C\": \"Overexposed\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001405", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the visual environment? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low light\", \"C\": \"Overexposed\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001405", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001406", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001406", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Windy\", \"C\": \"Thunderstorm\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001407", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Windy\", \"C\": \"Thunderstorm\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001407", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Pedestrian signal\", \"C\": \"None\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001408", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Pedestrian signal\", \"C\": \"None\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001408", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Potholes\", \"C\": \"Cracked asphalt\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001409", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/936.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Potholes\", \"C\": \"Cracked asphalt\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001409", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001410", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001410", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Backlit\", \"C\": \"Dawn\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001411", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Backlit\", \"C\": \"Dawn\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001411", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Congested\", \"C\": \"Heavy\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001412", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Congested\", \"C\": \"Heavy\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001412", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Mist\", \"C\": \"Clear\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001413", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Mist\", \"C\": \"Clear\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001413", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from this image? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Blinking green\", \"C\": \"Unknown\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001414", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from this image? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Blinking green\", \"C\": \"Unknown\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001414", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Debris scattered\", \"C\": \"Potholes\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001415", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/945.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Debris scattered\", \"C\": \"Potholes\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001415", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Harbor area\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001416", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Subway entrance\", \"C\": \"Harbor area\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001416", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene described? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Warm\", \"C\": \"Foggy\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001417", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene described? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Warm\", \"C\": \"Foggy\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001417", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Detour traffic\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001418", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Detour traffic\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001418", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Smoggy\", \"C\": \"Thunderstorm\", \"D\": \"Drizzling\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001419", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Smoggy\", \"C\": \"Thunderstorm\", \"D\": \"Drizzling\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001419", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Not visible\", \"C\": \"Covered by tree\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001420", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Not visible\", \"C\": \"Covered by tree\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001420", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Dry\", \"C\": \"Cracked asphalt\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001421", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/954.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Dry\", \"C\": \"Cracked asphalt\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001421", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Urban\", \"C\": \"Roundabout\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001422", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Urban\", \"C\": \"Roundabout\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001422", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Moderate\", \"C\": \"Foggy\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001423", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Moderate\", \"C\": \"Foggy\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001423", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001424", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001424", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dusty\", \"C\": \"Snowstorm\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001425", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dusty\", \"C\": \"Snowstorm\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001425", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Absent\", \"C\": \"Flashing red\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001426", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Absent\", \"C\": \"Flashing red\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001426", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001427", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/009_val_motion_blur/97.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001427", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Parkway\", \"B\": \"Suburban\", \"C\": \"Coastal road\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001428", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Parkway\", \"B\": \"Suburban\", \"C\": \"Coastal road\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Parkway" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001428", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Under tunnel\", \"C\": \"Soft light\", \"D\": \"Glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001429", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Under tunnel\", \"C\": \"Soft light\", \"D\": \"Glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Soft light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001429", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001430", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001430", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Clear\", \"C\": \"Mist\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001431", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Clear\", \"C\": \"Mist\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001431", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Off / malfunctioning\", \"C\": \"Green\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001432", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Off / malfunctioning\", \"C\": \"Green\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001432", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Muddy\", \"C\": \"Wet\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001433", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/0.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Muddy\", \"C\": \"Wet\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001433", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001434", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001434", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Backlit\", \"C\": \"Clear\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001435", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Backlit\", \"C\": \"Clear\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001435", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001436", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001436", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any adverse weather in the scene? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Gloomy\", \"C\": \"Flurries\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001437", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any adverse weather in the scene? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Gloomy\", \"C\": \"Flurries\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001437", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Temporary road sign light\", \"C\": \"Flashing amber\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001438", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Temporary road sign light\", \"C\": \"Flashing amber\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001438", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Potholes\", \"C\": \"Icy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001439", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/107.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Potholes\", \"C\": \"Icy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001439", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Two-lane\", \"B\": \"Service station\", \"C\": \"Parking lot\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001440", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Two-lane\", \"B\": \"Service station\", \"C\": \"Parking lot\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Two-lane" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001440", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in the scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Rain-blurred\", \"C\": \"Under tunnel\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001441", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in the scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Rain-blurred\", \"C\": \"Under tunnel\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001441", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition on the road? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001442", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition on the road? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001442", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Clear\", \"C\": \"Sleet\", \"D\": \"Thunderstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001443", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Clear\", \"C\": \"Sleet\", \"D\": \"Thunderstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001443", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Flashing amber\", \"C\": \"Pedestrian signal\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001444", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Flashing amber\", \"C\": \"Pedestrian signal\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001444", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Flooded\", \"C\": \"Gravel\", \"D\": \"Leaf-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001445", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/116.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Flooded\", \"C\": \"Gravel\", \"D\": \"Leaf-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001445", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Urban\", \"C\": \"Harbor area\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001446", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Urban\", \"C\": \"Harbor area\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001446", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001447", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001447", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition on the road? You should answer one choice from four answers \"A\": \"Empty\", \"B\": \"Lane merging\", \"C\": \"Detour traffic\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001448", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition on the road? You should answer one choice from four answers \"A\": \"Empty\", \"B\": \"Lane merging\", \"C\": \"Detour traffic\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Empty" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001448", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dry\", \"C\": \"Flurries\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001449", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dry\", \"C\": \"Flurries\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001449", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Yellow\", \"C\": \"Absent\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001450", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Yellow\", \"C\": \"Absent\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001450", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Partially damaged\", \"C\": \"Dry\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001451", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/125.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Partially damaged\", \"C\": \"Dry\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001451", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Roundabout\", \"C\": \"Urban\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001452", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Roundabout\", \"C\": \"Urban\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001452", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene described? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Partly cloudy\", \"C\": \"Overexposed\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001453", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene described? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Partly cloudy\", \"C\": \"Overexposed\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001453", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition on the road? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001454", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition on the road? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001454", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Strong wind\", \"C\": \"Mist\", \"D\": \"Overcast\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001455", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Strong wind\", \"C\": \"Mist\", \"D\": \"Overcast\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001455", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Red\", \"C\": \"No\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001456", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Red\", \"C\": \"No\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001456", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Dry\", \"C\": \"Potholes\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001457", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/134.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Dry\", \"C\": \"Potholes\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001457", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Suburban\", \"C\": \"Parking lot\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001458", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Suburban\", \"C\": \"Parking lot\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001458", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Dim streetlight visibility\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001459", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Dim streetlight visibility\", \"C\": \"Clear\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001459", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001460", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001460", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Dry\", \"C\": \"Stormy\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001461", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Dry\", \"C\": \"Stormy\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001461", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Covered by tree\", \"C\": \"Not visible\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001462", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Covered by tree\", \"C\": \"Not visible\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001462", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Snowy\", \"C\": \"Smooth\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001463", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/143.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Snowy\", \"C\": \"Smooth\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001463", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Service station\", \"C\": \"Rural\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001464", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Service station\", \"C\": \"Rural\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001464", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Clear\", \"C\": \"Glare\", \"D\": \"Reflected glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001465", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Clear\", \"C\": \"Glare\", \"D\": \"Reflected glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001465", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001466", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001466", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Hazy\", \"C\": \"Mist\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001467", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Hazy\", \"C\": \"Mist\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001467", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Dim / unclear\", \"C\": \"None\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001468", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Dim / unclear\", \"C\": \"None\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001468", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001469", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/152.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Flooded\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001469", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Campus area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001470", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Campus area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001470", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting condition in this scene? You should answer one choice from four answers \"A\": \"Warm light\", \"B\": \"Overexposed\", \"C\": \"Foggy\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001471", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting condition in this scene? You should answer one choice from four answers \"A\": \"Warm light\", \"B\": \"Overexposed\", \"C\": \"Foggy\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001471", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there active traffic on the road? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001472", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there active traffic on the road? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Detour traffic\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001472", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Dusty\", \"C\": \"Dry\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001473", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Dusty\", \"C\": \"Dry\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001473", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Reflected signal\", \"C\": \"None\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001474", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Reflected signal\", \"C\": \"None\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001474", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Icy\", \"C\": \"Dry\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001475", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/161.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Icy\", \"C\": \"Dry\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001475", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Rural\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001476", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Rural\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001476", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Dawn\", \"C\": \"Dim streetlight visibility\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001477", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Dawn\", \"C\": \"Dim streetlight visibility\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001477", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001478", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001478", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Snowy\", \"C\": \"Dry\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001479", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Snowy\", \"C\": \"Dry\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001479", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Yellow\", \"C\": \"Not visible\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001480", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Yellow\", \"C\": \"Not visible\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001480", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Smooth\", \"C\": \"Cracked asphalt\", \"D\": \"Uneven\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001481", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/170.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Smooth\", \"C\": \"Cracked asphalt\", \"D\": \"Uneven\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001481", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of street is depicted in the scene? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Harbor area\", \"C\": \"Tree-lined street\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001482", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of street is depicted in the scene? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Harbor area\", \"C\": \"Tree-lined street\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Tree-lined street" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001482", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Backlit\", \"C\": \"Sunny\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001483", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Backlit\", \"C\": \"Sunny\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001483", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Lane merging\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001484", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Detour traffic\", \"C\": \"Lane merging\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001484", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Dusty\", \"C\": \"Sunny\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001485", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Dusty\", \"C\": \"Sunny\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001485", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Dim / unclear\", \"C\": \"Temporary road sign light\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001486", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Absent\", \"B\": \"Dim / unclear\", \"C\": \"Temporary road sign light\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001486", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Sandy\", \"C\": \"Potholes\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001487", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/18.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Sandy\", \"C\": \"Potholes\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001487", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Urban\", \"C\": \"Bridge\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001488", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Urban\", \"C\": \"Bridge\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001488", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Clear\", \"C\": \"Under tunnel\", \"D\": \"Glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001489", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Clear\", \"C\": \"Under tunnel\", \"D\": \"Glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001489", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001490", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001490", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snowstorm\", \"C\": \"Mist\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001491", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snowstorm\", \"C\": \"Mist\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001491", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Blinking green\", \"C\": \"Covered by tree\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001492", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Blinking green\", \"C\": \"Covered by tree\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001492", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Smooth\", \"C\": \"Water puddles\", \"D\": \"Uneven\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001493", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/189.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Smooth\", \"C\": \"Water puddles\", \"D\": \"Uneven\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001493", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Harbor area\", \"D\": \"Construction zone\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001494", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Harbor area\", \"D\": \"Construction zone\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001494", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene described? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Hazy\", \"C\": \"Reflected glare\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001495", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene described? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Hazy\", \"C\": \"Reflected glare\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001495", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition on the road? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Heavy\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001496", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition on the road? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Heavy\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001496", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Flurries\", \"C\": \"Clear\", \"D\": \"Windy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001497", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Flurries\", \"C\": \"Clear\", \"D\": \"Windy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001497", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the frame? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Countdown timer visible\", \"C\": \"Off / malfunctioning\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001498", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the frame? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Countdown timer visible\", \"C\": \"Off / malfunctioning\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001498", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Wet\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001499", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/198.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Wet\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001499", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Urban\", \"C\": \"Subway entrance\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001500", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Urban\", \"C\": \"Subway entrance\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001500", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Rain-blurred\", \"C\": \"Clear\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001501", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Rain-blurred\", \"C\": \"Clear\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001501", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001502", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001502", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Hailing\", \"C\": \"Drizzling\", \"D\": \"Haze\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001503", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Hailing\", \"C\": \"Drizzling\", \"D\": \"Haze\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001503", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Green\", \"C\": \"Not visible\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001504", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Green\", \"C\": \"Not visible\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001504", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Leaf-covered\", \"C\": \"Wet\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001505", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/206.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Leaf-covered\", \"C\": \"Wet\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001505", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001506", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001506", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Low\", \"C\": \"Backlit\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001507", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Low\", \"C\": \"Backlit\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001507", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this view? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001508", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this view? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001508", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Rainy\", \"C\": \"Windy\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001509", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Rainy\", \"C\": \"Windy\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001509", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Blinking green\", \"C\": \"Countdown timer visible\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001510", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Blinking green\", \"C\": \"Countdown timer visible\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001510", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Construction debris\", \"C\": \"Potholes\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001511", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/215.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Construction debris\", \"C\": \"Potholes\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001511", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Residential\", \"C\": \"Coastal road\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001512", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Residential\", \"C\": \"Coastal road\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001512", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dawn\", \"C\": \"Under tunnel\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001513", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Dawn\", \"C\": \"Under tunnel\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001513", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001514", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001514", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Sleet\", \"C\": \"Clear\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001515", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Sleet\", \"C\": \"Clear\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001515", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Flashing amber\", \"C\": \"Unknown\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001516", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Flashing amber\", \"C\": \"Unknown\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001516", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Snowy\", \"C\": \"Dry\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001517", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/224.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Snowy\", \"C\": \"Dry\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001517", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Coastal road\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001518", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Roundabout\", \"C\": \"Coastal road\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001518", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Blinding sunlight\", \"C\": \"Clear\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001519", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Blinding sunlight\", \"C\": \"Clear\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001519", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001520", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001520", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Strong wind\", \"C\": \"Clear\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001521", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Strong wind\", \"C\": \"Clear\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001521", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Blinking green\", \"C\": \"None\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001522", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Blinking green\", \"C\": \"None\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001522", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Rough surface\", \"C\": \"Dry\", \"D\": \"Icy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001523", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/233.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Rough surface\", \"C\": \"Dry\", \"D\": \"Icy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001523", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of street is depicted in the scene? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Harbor area\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001524", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of street is depicted in the scene? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Harbor area\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001524", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Blinding sunlight\", \"C\": \"Dim streetlight visibility\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001525", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Blinding sunlight\", \"C\": \"Dim streetlight visibility\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001525", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this area? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001526", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this area? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001526", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Sunny\", \"C\": \"Windy\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001527", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Sunny\", \"C\": \"Windy\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001527", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Not visible\", \"C\": \"Flashing amber\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001528", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Not visible\", \"C\": \"Flashing amber\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001528", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Uneven\", \"C\": \"Newly paved\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001529", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/242.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Uneven\", \"C\": \"Newly paved\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001529", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Rural\", \"C\": \"Harbor area\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001530", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Rural\", \"C\": \"Harbor area\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001530", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Clear\", \"C\": \"Low light\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001531", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Clear\", \"C\": \"Low light\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001531", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001532", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001532", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Drizzling\", \"C\": \"Foggy\", \"D\": \"Dusty\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001533", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Drizzling\", \"C\": \"Foggy\", \"D\": \"Dusty\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001533", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Not visible\", \"C\": \"Covered by tree\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001534", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Not visible\", \"C\": \"Covered by tree\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001534", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Dry\", \"C\": \"Bumpy\", \"D\": \"Icy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001535", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/251.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Dry\", \"C\": \"Bumpy\", \"D\": \"Icy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001535", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the image? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001536", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the image? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001536", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Glare\", \"C\": \"Rain-blurred\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001537", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Glare\", \"C\": \"Rain-blurred\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001537", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in the scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001538", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in the scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001538", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Drizzling\", \"C\": \"Overcast\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001539", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Drizzling\", \"C\": \"Overcast\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001539", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Off / malfunctioning\", \"C\": \"Unknown\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001540", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Off / malfunctioning\", \"C\": \"Unknown\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001540", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Debris scattered\", \"C\": \"Gravel\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001541", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/260.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Debris scattered\", \"C\": \"Gravel\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001541", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Parking lot\", \"C\": \"Harbor area\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001542", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Parking lot\", \"C\": \"Harbor area\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001542", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001543", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001543", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001544", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001544", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Flurries\", \"C\": \"Snowstorm\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001545", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Flurries\", \"C\": \"Snowstorm\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001545", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic signal visible? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Not visible\", \"C\": \"Countdown timer visible\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001546", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is a traffic signal visible? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Not visible\", \"C\": \"Countdown timer visible\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001546", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Sandy\", \"C\": \"Gravel\", \"D\": \"Debris scattered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001547", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/27.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Sandy\", \"C\": \"Gravel\", \"D\": \"Debris scattered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001547", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Forest road\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001548", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this street? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Forest road\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001548", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Low light\", \"C\": \"Under tunnel\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001549", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Low light\", \"C\": \"Under tunnel\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001549", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this area? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001550", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this area? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001550", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Sandstorm\", \"C\": \"Rainy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001551", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Sandstorm\", \"C\": \"Rainy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001551", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Pedestrian signal\", \"C\": \"Blinking green\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001552", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Pedestrian signal\", \"C\": \"Blinking green\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001552", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Snowy\", \"C\": \"Dry\", \"D\": \"Cracked asphalt\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001553", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/279.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Snowy\", \"C\": \"Dry\", \"D\": \"Cracked asphalt\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001553", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Urban\", \"C\": \"Rural\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001554", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Urban\", \"C\": \"Rural\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001554", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Under tunnel\", \"C\": \"Foggy\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001555", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Under tunnel\", \"C\": \"Foggy\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001555", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001556", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001556", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Sleet\", \"C\": \"Foggy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001557", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Sleet\", \"C\": \"Foggy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001557", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Flashing yellow\", \"C\": \"Not visible\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001558", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Flashing yellow\", \"C\": \"Not visible\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001558", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Cracked asphalt\", \"C\": \"Flooded\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001559", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/288.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Cracked asphalt\", \"C\": \"Flooded\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001559", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Highway\", \"C\": \"Service station\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001560", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Highway\", \"C\": \"Service station\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001560", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Dim streetlight visibility\", \"C\": \"Backlit\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001561", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Dim streetlight visibility\", \"C\": \"Backlit\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001561", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001562", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001562", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Windy\", \"C\": \"Dusty\", \"D\": \"Mist\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001563", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Windy\", \"C\": \"Dusty\", \"D\": \"Mist\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001563", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Temporary road sign light\", \"C\": \"Not visible\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001564", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Temporary road sign light\", \"C\": \"Not visible\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001564", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Gravel\", \"D\": \"Uneven\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001565", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/297.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Gravel\", \"D\": \"Uneven\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001565", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001566", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001566", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Blurry\", \"B\": \"Foggy\", \"C\": \"Reflected glare\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001567", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Blurry\", \"B\": \"Foggy\", \"C\": \"Reflected glare\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Blurry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001567", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Detour traffic\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001568", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Detour traffic\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001568", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Strong wind\", \"C\": \"Rainy\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001569", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Strong wind\", \"C\": \"Rainy\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001569", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Green\", \"C\": \"Absent\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001570", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Green\", \"C\": \"Absent\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001570", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Cracked asphalt\", \"C\": \"Rough surface\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001571", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/305.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Cracked asphalt\", \"C\": \"Rough surface\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001571", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001572", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001572", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Strong shadows\", \"C\": \"Dusk\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001573", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Strong shadows\", \"C\": \"Dusk\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001573", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001574", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001574", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Snowstorm\", \"C\": \"Haze\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001575", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Windy\", \"B\": \"Snowstorm\", \"C\": \"Haze\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001575", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"None\", \"C\": \"Dim / unclear\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001576", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"None\", \"C\": \"Dim / unclear\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001576", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Blocked by obstacle\", \"C\": \"Snowy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001577", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/314.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Blocked by obstacle\", \"C\": \"Snowy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001577", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Urban\", \"C\": \"Airport roadway\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001578", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Urban\", \"C\": \"Airport roadway\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001578", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Partly cloudy\", \"C\": \"Flashing lights\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001579", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Partly cloudy\", \"C\": \"Flashing lights\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001579", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001580", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001580", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Cloudy\", \"C\": \"Thunderstorm\", \"D\": \"Mist\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001581", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Cloudy\", \"C\": \"Thunderstorm\", \"D\": \"Mist\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001581", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Green\", \"C\": \"Red\", \"D\": \"Absent\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001582", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Green\", \"C\": \"Red\", \"D\": \"Absent\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001582", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Wet\", \"C\": \"Dry\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001583", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/323.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Wet\", \"C\": \"Dry\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001583", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Airport roadway\", \"C\": \"Forest road\", \"D\": \"Campus area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001584", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Airport roadway\", \"C\": \"Forest road\", \"D\": \"Campus area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001584", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Dusk\", \"C\": \"Clear\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001585", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Partly cloudy\", \"B\": \"Dusk\", \"C\": \"Clear\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001585", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001586", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001586", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Snowy\", \"C\": \"Sandstorm\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001587", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Snowy\", \"C\": \"Sandstorm\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001587", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Off / malfunctioning\", \"C\": \"Yellow\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001588", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Off / malfunctioning\", \"C\": \"Yellow\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001588", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Gravel\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001589", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/332.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Gravel\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001589", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001590", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001590", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Hazy\", \"C\": \"Dusk\", \"D\": \"Limited\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001591", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Hazy\", \"C\": \"Dusk\", \"D\": \"Limited\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Limited" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001591", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001592", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001592", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Haze\", \"C\": \"Windy\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001593", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Haze\", \"C\": \"Windy\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001593", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Blinking green\", \"C\": \"Reflected signal\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001594", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Blinking green\", \"C\": \"Reflected signal\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001594", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Dry\", \"C\": \"Muddy\", \"D\": \"Icy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001595", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/341.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Dry\", \"C\": \"Muddy\", \"D\": \"Icy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001595", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Urban\", \"C\": \"Service station\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001596", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Urban\", \"C\": \"Service station\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001596", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Hazy\", \"C\": \"Warm\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001597", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Hazy\", \"C\": \"Warm\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001597", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Rush hour\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001598", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Rush hour\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001598", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Strong wind\", \"C\": \"Clear\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001599", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Strong wind\", \"C\": \"Clear\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001599", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Green\", \"C\": \"Absent\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001600", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Green\", \"C\": \"Absent\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001600", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snow-covered\", \"C\": \"Cracked asphalt\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001601", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/350.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snow-covered\", \"C\": \"Cracked asphalt\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001601", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Campus area\", \"C\": \"Urban\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001602", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Airport roadway\", \"B\": \"Campus area\", \"C\": \"Urban\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001602", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Rain-blurred\", \"C\": \"Hazy\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001603", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Rain-blurred\", \"C\": \"Hazy\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001603", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle traffic in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001604", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle traffic in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001604", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Hazy\", \"C\": \"Snowy\", \"D\": \"Windy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001605", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Hazy\", \"C\": \"Snowy\", \"D\": \"Windy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001605", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in this scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Covered by tree\", \"C\": \"No\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001606", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in this scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Covered by tree\", \"C\": \"No\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001606", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Dry\", \"C\": \"Water puddles\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001607", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/36.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Dry\", \"C\": \"Water puddles\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001607", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001608", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001608", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Hazy\", \"C\": \"Flashing lights\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001609", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Hazy\", \"C\": \"Flashing lights\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001609", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Detour traffic\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001610", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Congested\", \"C\": \"Detour traffic\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001610", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Smoggy\", \"C\": \"Haze\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001611", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Smoggy\", \"C\": \"Haze\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001611", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Not visible\", \"C\": \"Countdown timer visible\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001612", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Not visible\", \"C\": \"Countdown timer visible\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001612", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Uneven\", \"C\": \"Newly paved\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001613", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/369.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Uneven\", \"C\": \"Newly paved\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001613", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001614", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001614", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Foggy\", \"C\": \"Hazy\", \"D\": \"Strong shadows\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001615", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Foggy\", \"C\": \"Hazy\", \"D\": \"Strong shadows\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001615", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001616", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001616", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Flurries\", \"C\": \"Clear\", \"D\": \"Mist\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001617", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Flurries\", \"C\": \"Clear\", \"D\": \"Mist\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001617", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"None\", \"C\": \"Yellow\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001618", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"None\", \"C\": \"Yellow\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001618", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Snowy\", \"C\": \"Dry\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001619", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/378.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Snowy\", \"C\": \"Dry\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001619", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Shopping district\", \"C\": \"Urban\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001620", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Shopping district\", \"C\": \"Urban\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001620", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Hazy\", \"C\": \"Dim streetlight visibility\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001621", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Rain-blurred\", \"B\": \"Hazy\", \"C\": \"Dim streetlight visibility\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001621", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Detour traffic\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001622", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Detour traffic\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001622", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Gloomy\", \"C\": \"Overcast\", \"D\": \"Haze\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001623", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Dusty\", \"B\": \"Gloomy\", \"C\": \"Overcast\", \"D\": \"Haze\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001623", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Not visible\", \"C\": \"Green\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001624", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Not visible\", \"C\": \"Green\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001624", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Snow-covered\", \"C\": \"Leaf-covered\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001625", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/387.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Snow-covered\", \"C\": \"Leaf-covered\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001625", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Bridge\", \"C\": \"Forest road\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001626", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Bridge\", \"C\": \"Forest road\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001626", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Reflected glare\", \"C\": \"Golden\", \"D\": \"Under tunnel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001627", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Reflected glare\", \"C\": \"Golden\", \"D\": \"Under tunnel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Golden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001627", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001628", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Congested\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001628", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Hailing\", \"C\": \"Rainy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001629", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Hailing\", \"C\": \"Rainy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001629", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Unknown\", \"C\": \"Covered by tree\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001630", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Unknown\", \"C\": \"Covered by tree\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001630", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Dry\", \"C\": \"Rough surface\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001631", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/396.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Dry\", \"C\": \"Rough surface\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001631", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001632", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001632", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dusk\", \"C\": \"Clear\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001633", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dusk\", \"C\": \"Clear\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001633", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001634", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001634", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Gloomy\", \"C\": \"Sleet\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001635", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Gloomy\", \"C\": \"Sleet\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001635", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Temporary road sign light\", \"C\": \"Not visible\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001636", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Temporary road sign light\", \"C\": \"Not visible\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001636", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Blocked by obstacle\", \"C\": \"Dry\", \"D\": \"Icy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001637", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/404.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Blocked by obstacle\", \"C\": \"Dry\", \"D\": \"Icy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001637", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Forest road\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001638", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Forest road\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001638", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Moderate\", \"C\": \"Low light\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001639", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Moderate\", \"C\": \"Low light\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001639", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001640", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001640", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Sandstorm\", \"C\": \"Rainy\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001641", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Sandstorm\", \"C\": \"Rainy\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001641", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Absent\", \"C\": \"Countdown timer visible\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001642", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Absent\", \"C\": \"Countdown timer visible\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001642", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Leaf-covered\", \"C\": \"Muddy\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001643", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/413.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Leaf-covered\", \"C\": \"Muddy\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001643", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Urban\", \"C\": \"Campus area\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001644", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Urban\", \"C\": \"Campus area\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001644", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Partly cloudy\", \"C\": \"Reflected glare\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001645", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Partly cloudy\", \"C\": \"Reflected glare\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001645", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001646", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001646", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Strong wind\", \"C\": \"Sleet\", \"D\": \"Haze\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001647", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Strong wind\", \"C\": \"Sleet\", \"D\": \"Haze\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001647", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"No\", \"C\": \"Countdown timer visible\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001648", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"No\", \"C\": \"Countdown timer visible\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001648", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Wet\", \"C\": \"Newly paved\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001649", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/422.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Wet\", \"C\": \"Newly paved\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001649", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Urban\", \"C\": \"Suburban\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001650", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Urban\", \"C\": \"Suburban\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001650", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Clear\", \"C\": \"Partly cloudy\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001651", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Clear\", \"C\": \"Partly cloudy\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001651", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Lane merging\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001652", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Lane merging\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001652", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Smoggy\", \"C\": \"Dry\", \"D\": \"Snowstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001653", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Hailing\", \"B\": \"Smoggy\", \"C\": \"Dry\", \"D\": \"Snowstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001653", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Hidden\", \"C\": \"Red\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001654", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Hidden\", \"C\": \"Red\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001654", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Dry\", \"C\": \"Rough surface\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001655", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/431.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Dry\", \"C\": \"Rough surface\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001655", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Roundabout\", \"C\": \"Campus area\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001656", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Roundabout\", \"C\": \"Campus area\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001656", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in the scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Hazy\", \"C\": \"Rain-blurred\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001657", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in the scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Hazy\", \"C\": \"Rain-blurred\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001657", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001658", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Accident blockage\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001658", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Haze\", \"C\": \"Windy\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001659", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Haze\", \"C\": \"Windy\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001659", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the frame? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Red\", \"C\": \"Pedestrian signal\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001660", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the frame? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Red\", \"C\": \"Pedestrian signal\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001660", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Dry\", \"C\": \"Leaf-covered\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001661", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/440.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Dry\", \"C\": \"Leaf-covered\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001661", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Rural\", \"C\": \"Forest road\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001662", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Rural\", \"C\": \"Forest road\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001662", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Clear\", \"C\": \"Dawn\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001663", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Clear\", \"C\": \"Dawn\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001663", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001664", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001664", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Clear\", \"C\": \"Drizzling\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001665", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Clear\", \"C\": \"Drizzling\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001665", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Reflected signal\", \"C\": \"Unknown\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001666", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Reflected signal\", \"C\": \"Unknown\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001666", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Dry\", \"C\": \"Muddy\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001667", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/45.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Dry\", \"C\": \"Muddy\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001667", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001668", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001668", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Blurry\", \"C\": \"Smoky\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001669", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Blurry\", \"C\": \"Smoky\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Blurry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001669", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Rush hour\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001670", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Rush hour\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001670", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Hazy\", \"C\": \"Drizzling\", \"D\": \"Strong wind\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001671", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Hazy\", \"C\": \"Drizzling\", \"D\": \"Strong wind\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001671", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Absent\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001672", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Absent\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001672", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Potholes\", \"C\": \"Rough surface\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001673", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/459.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Potholes\", \"C\": \"Rough surface\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001673", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Coastal road\", \"C\": \"Forest road\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001674", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Coastal road\", \"C\": \"Forest road\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001674", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting quality described? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Rain-blurred\", \"C\": \"Hazy\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001675", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting quality described? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Rain-blurred\", \"C\": \"Hazy\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001675", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001676", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001676", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any visible weather condition affecting the scene? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Gloomy\", \"C\": \"Thunderstorm\", \"D\": \"Haze\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001677", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any visible weather condition affecting the scene? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Gloomy\", \"C\": \"Thunderstorm\", \"D\": \"Haze\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001677", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Flashing red\", \"C\": \"Absent\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001678", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Reflected signal\", \"B\": \"Flashing red\", \"C\": \"Absent\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001678", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Dry\", \"C\": \"Snow-covered\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001679", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/468.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Dry\", \"C\": \"Snow-covered\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001679", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Airport roadway\", \"C\": \"Suburban\", \"D\": \"Campus area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001680", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Airport roadway\", \"C\": \"Suburban\", \"D\": \"Campus area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001680", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Low light\", \"C\": \"Clear\", \"D\": \"Dusk\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001681", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Low light\", \"C\": \"Clear\", \"D\": \"Dusk\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001681", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001682", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001682", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Gloomy\", \"C\": \"Cloudy\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001683", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Gloomy\", \"C\": \"Cloudy\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001683", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001684", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001684", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Debris scattered\", \"C\": \"Smooth\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001685", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/477.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Debris scattered\", \"C\": \"Smooth\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001685", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Construction zone\", \"C\": \"Suburban\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001686", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Construction zone\", \"C\": \"Suburban\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001686", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Flashing lights\", \"C\": \"Overexposed\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001687", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Flashing lights\", \"C\": \"Overexposed\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001687", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic situation? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001688", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic situation? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001688", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Gloomy\", \"C\": \"Dry\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001689", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Gloomy\", \"C\": \"Dry\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001689", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"None\", \"C\": \"Flashing amber\", \"D\": \"Flashing yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001690", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"None\", \"C\": \"Flashing amber\", \"D\": \"Flashing yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001690", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Dry\", \"C\": \"Slippery\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001691", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/486.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Dry\", \"C\": \"Slippery\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001691", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Shopping district\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001692", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Forest road\", \"C\": \"Shopping district\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001692", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Backlit\", \"C\": \"Dim streetlight visibility\", \"D\": \"Warm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001693", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Backlit\", \"C\": \"Dim streetlight visibility\", \"D\": \"Warm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001693", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like on the road? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001694", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like on the road? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001694", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Dry\", \"C\": \"Hailing\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001695", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sleet\", \"B\": \"Dry\", \"C\": \"Hailing\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001695", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"No\", \"C\": \"Reflected signal\", \"D\": \"None\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001696", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"No\", \"C\": \"Reflected signal\", \"D\": \"None\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001696", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Snowy\", \"C\": \"Construction debris\", \"D\": \"Wet\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001697", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/495.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Snowy\", \"C\": \"Construction debris\", \"D\": \"Wet\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001697", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Forest road\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001698", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Forest road\", \"C\": \"Highway\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001698", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting quality in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Soft\", \"C\": \"Clear\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001699", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting quality in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Soft\", \"C\": \"Clear\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Soft" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001699", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001700", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001700", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Flurries\", \"C\": \"Drizzling\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001701", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Flurries\", \"C\": \"Drizzling\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001701", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Reflected signal\", \"C\": \"None\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001702", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Reflected signal\", \"C\": \"None\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001702", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001703", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/503.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Dry\", \"C\": \"Sandy\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001703", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Urban\", \"C\": \"Highway\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001704", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Suburban\", \"B\": \"Urban\", \"C\": \"Highway\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001704", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Partly cloudy\", \"C\": \"Excellent\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001705", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Partly cloudy\", \"C\": \"Excellent\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Excellent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001705", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001706", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001706", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Sunny\", \"C\": \"Rainy\", \"D\": \"Dusty\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001707", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Sunny\", \"C\": \"Rainy\", \"D\": \"Dusty\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001707", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Pedestrian signal\", \"C\": \"Flashing yellow\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001708", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Pedestrian signal\", \"C\": \"Flashing yellow\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001708", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Bumpy\", \"C\": \"Leaf-covered\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001709", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/512.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Bumpy\", \"C\": \"Leaf-covered\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001709", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is shown? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001710", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is shown? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001710", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Clear\", \"C\": \"Overexposed\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001711", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Clear\", \"C\": \"Overexposed\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001711", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001712", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001712", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Windy\", \"C\": \"Hailing\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001713", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Windy\", \"C\": \"Hailing\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001713", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"No\", \"C\": \"Off / malfunctioning\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001714", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"No\", \"C\": \"Off / malfunctioning\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001714", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Blocked by obstacle\", \"C\": \"Dry\", \"D\": \"Debris scattered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001715", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/521.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Blocked by obstacle\", \"C\": \"Dry\", \"D\": \"Debris scattered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001715", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Residential\", \"C\": \"Coastal road\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001716", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is this? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Residential\", \"C\": \"Coastal road\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001716", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Flashing lights\", \"C\": \"Reflected glare\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001717", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Flashing lights\", \"C\": \"Reflected glare\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001717", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001718", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Heavy\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001718", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Sunny\", \"C\": \"Sleet\", \"D\": \"Thunderstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001719", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Sunny\", \"C\": \"Sleet\", \"D\": \"Thunderstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001719", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Temporary road sign light\", \"C\": \"Unknown\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001720", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Temporary road sign light\", \"C\": \"Unknown\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001720", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Dry\", \"C\": \"Slippery\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001721", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/530.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Dry\", \"C\": \"Slippery\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001721", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Airport roadway\", \"C\": \"Suburban\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001722", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Airport roadway\", \"C\": \"Suburban\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001722", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Strong shadows\", \"C\": \"Dusk\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001723", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Strong shadows\", \"C\": \"Dusk\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001723", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Accident blockage\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001724", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Accident blockage\", \"C\": \"Detour traffic\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001724", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Sunny\", \"C\": \"Sandstorm\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001725", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Gloomy\", \"B\": \"Sunny\", \"C\": \"Sandstorm\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001725", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Reflected signal\", \"C\": \"Red\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001726", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Reflected signal\", \"C\": \"Red\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001726", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Bumpy\", \"C\": \"Dry\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001727", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/54.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Bumpy\", \"C\": \"Dry\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001727", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Subway entrance\", \"C\": \"Urban\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001728", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Subway entrance\", \"C\": \"Urban\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001728", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Rain-blurred\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001729", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Clear\", \"C\": \"Rain-blurred\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001729", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Congested\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001730", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Congested\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001730", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Clear\", \"C\": \"Stormy\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001731", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Clear\", \"C\": \"Stormy\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001731", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Off / malfunctioning\", \"C\": \"Dim / unclear\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001732", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Blinking green\", \"B\": \"Off / malfunctioning\", \"C\": \"Dim / unclear\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001732", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Snow-covered\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001733", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/549.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Dry\", \"C\": \"Snow-covered\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001733", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001734", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Service station\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001734", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Under tunnel\", \"C\": \"Clear\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001735", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Under tunnel\", \"C\": \"Clear\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001735", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition in the area? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001736", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition in the area? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001736", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Windy\", \"C\": \"Overcast\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001737", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Windy\", \"C\": \"Overcast\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001737", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Pedestrian signal\", \"C\": \"Flashing yellow\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001738", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Pedestrian signal\", \"C\": \"Flashing yellow\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001738", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Gravel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001739", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/558.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Gravel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001739", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001740", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Service station\", \"B\": \"Harbor area\", \"C\": \"Urban\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001740", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Glare\", \"C\": \"Under tunnel\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001741", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Glare\", \"C\": \"Under tunnel\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001741", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Lane merging\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001742", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Lane merging\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001742", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Rainy\", \"C\": \"Clear\", \"D\": \"Strong wind\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001743", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Rainy\", \"C\": \"Clear\", \"D\": \"Strong wind\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001743", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Off / malfunctioning\", \"C\": \"Flashing amber\", \"D\": \"Hidden\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001744", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible? You should answer one choice from four answers \"A\": \"Green\", \"B\": \"Off / malfunctioning\", \"C\": \"Flashing amber\", \"D\": \"Hidden\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001744", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Dry\", \"C\": \"Partially damaged\", \"D\": \"Oil slick\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001745", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/567.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Dry\", \"C\": \"Partially damaged\", \"D\": \"Oil slick\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001745", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001746", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Urban\", \"C\": \"Coastal road\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001746", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the visual environment? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Foggy\", \"C\": \"Smoky\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001747", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the visual environment? You should answer one choice from four answers \"A\": \"Flashing lights\", \"B\": \"Foggy\", \"C\": \"Smoky\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001747", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001748", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001748", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather like? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Sunny\", \"C\": \"Sandstorm\", \"D\": \"Stormy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001749", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather like? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Sunny\", \"C\": \"Sandstorm\", \"D\": \"Stormy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001749", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Temporary road sign light\", \"C\": \"Hidden\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001750", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Temporary road sign light\", \"C\": \"Hidden\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001750", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Partially damaged\", \"C\": \"Dry\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001751", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/576.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Blocked by obstacle\", \"B\": \"Partially damaged\", \"C\": \"Dry\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001751", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001752", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Suburban\", \"C\": \"Urban\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001752", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Backlit\", \"C\": \"Glare\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001753", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Backlit\", \"C\": \"Glare\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001753", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001754", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Heavy\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001754", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Clear\", \"C\": \"Rainy\", \"D\": \"Snowstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001755", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Clear\", \"C\": \"Rainy\", \"D\": \"Snowstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001755", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"None\", \"C\": \"Not visible\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001756", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"None\", \"C\": \"Not visible\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001756", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Potholes\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001757", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/585.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Potholes\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001757", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road setting? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Shopping district\", \"C\": \"Airport roadway\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001758", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road setting? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Shopping district\", \"C\": \"Airport roadway\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001758", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visual clarity in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Dusk\", \"C\": \"Clear\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001759", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visual clarity in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Dusk\", \"C\": \"Clear\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001759", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001760", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Moderate\", \"C\": \"Detour traffic\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001760", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Dusty\", \"C\": \"Sunny\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001761", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Dusty\", \"C\": \"Sunny\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001761", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Hidden\", \"B\": \"Yellow\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001762", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Hidden\", \"B\": \"Yellow\", \"C\": \"Pedestrian signal\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001762", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001763", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/594.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Rough surface\", \"B\": \"Leaf-covered\", \"C\": \"Dry\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001763", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Roundabout\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001764", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Roundabout\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001764", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Blinding sunlight\", \"C\": \"Backlit\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001765", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Blinding sunlight\", \"C\": \"Backlit\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001765", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001766", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Accident blockage\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001766", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Dusty\", \"C\": \"Clear\", \"D\": \"Strong wind\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001767", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Dusty\", \"C\": \"Clear\", \"D\": \"Strong wind\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001767", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Yellow\", \"C\": \"No\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001768", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Yellow\", \"C\": \"No\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001768", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Rough surface\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001769", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/602.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Construction debris\", \"C\": \"Dry\", \"D\": \"Rough surface\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001769", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Parking lot\", \"C\": \"Harbor area\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001770", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Parking lot\", \"C\": \"Harbor area\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001770", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001771", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How clear is the visibility in this scene? You should answer one choice from four answers \"A\": \"Reflected glare\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001771", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001772", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Detour traffic\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001772", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Sandstorm\", \"C\": \"Sleet\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001773", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowstorm\", \"B\": \"Sandstorm\", \"C\": \"Sleet\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001773", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Red\", \"C\": \"Unknown\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001774", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Red\", \"C\": \"Unknown\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001774", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Potholes\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001775", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/611.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Leaf-covered\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Potholes\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001775", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001776", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001776", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Under tunnel\", \"C\": \"Blinding sunlight\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001777", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Under tunnel\", \"C\": \"Blinding sunlight\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001777", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001778", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Lane merging\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001778", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Flurries\", \"C\": \"Clear\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001779", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Overcast\", \"B\": \"Flurries\", \"C\": \"Clear\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001779", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Not visible\", \"C\": \"None\", \"D\": \"Red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001780", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Not visible\", \"C\": \"None\", \"D\": \"Red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001780", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Dry\", \"C\": \"Cracked asphalt\", \"D\": \"Sandy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001781", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/620.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Dry\", \"C\": \"Cracked asphalt\", \"D\": \"Sandy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001781", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Coastal road\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001782", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Coastal road\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001782", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Clear\", \"C\": \"Rain-blurred\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001783", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Clear\", \"C\": \"Rain-blurred\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001783", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic situation on the road? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Rush hour\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001784", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic situation on the road? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Rush hour\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001784", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Thunderstorm\", \"C\": \"Rainy\", \"D\": \"Drizzling\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001785", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Thunderstorm\", \"C\": \"Rainy\", \"D\": \"Drizzling\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001785", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Green\", \"C\": \"Flashing amber\", \"D\": \"Absent\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001786", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"Pedestrian signal\", \"B\": \"Green\", \"C\": \"Flashing amber\", \"D\": \"Absent\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001786", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Dry\", \"C\": \"Flooded\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001787", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/63.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Dry\", \"C\": \"Flooded\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001787", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Coastal road\", \"C\": \"Subway entrance\", \"D\": \"Residential\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001788", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Coastal road\", \"C\": \"Subway entrance\", \"D\": \"Residential\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001788", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting quality described in the scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Clear\", \"C\": \"Overexposed\", \"D\": \"Glare\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001789", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting quality described in the scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Clear\", \"C\": \"Overexposed\", \"D\": \"Glare\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Overexposed" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001789", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle activity in the scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001790", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle activity in the scene? You should answer one choice from four answers \"A\": \"Accident blockage\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001790", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Haze\", \"C\": \"Sleet\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001791", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Flurries\", \"B\": \"Haze\", \"C\": \"Sleet\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001791", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"None\", \"C\": \"Absent\", \"D\": \"Yellow\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001792", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"None\", \"C\": \"Absent\", \"D\": \"Yellow\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001792", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Potholes\", \"C\": \"Dry\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001793", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/639.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Gravel\", \"B\": \"Potholes\", \"C\": \"Dry\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001793", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Suburban\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001794", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Campus area\", \"C\": \"Suburban\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001794", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Clear\", \"C\": \"Blinding sunlight\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001795", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Clear\", \"C\": \"Blinding sunlight\", \"D\": \"Dim streetlight visibility\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001795", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Congested\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001796", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Lane merging\", \"C\": \"Congested\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001796", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Dusty\", \"C\": \"Dry\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001797", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Dusty\", \"C\": \"Dry\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001797", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Covered by tree\", \"C\": \"Not visible\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001798", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Covered by tree\", \"C\": \"Not visible\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001798", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Oil slick\", \"C\": \"Dry\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001799", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/648.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Oil slick\", \"C\": \"Dry\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001799", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Coastal road\", \"C\": \"Roundabout\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001800", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Coastal road\", \"C\": \"Roundabout\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001800", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Reflected glare\", \"C\": \"Clear\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001801", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Reflected glare\", \"C\": \"Clear\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001801", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001802", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001802", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Snowy\", \"C\": \"Sandstorm\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001803", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Snowy\", \"C\": \"Sandstorm\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001803", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Not visible\", \"C\": \"Blinking green\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001804", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Not visible\", \"C\": \"Blinking green\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001804", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Cracked asphalt\", \"C\": \"Smooth\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001805", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/657.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Cracked asphalt\", \"C\": \"Smooth\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001805", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Urban\", \"C\": \"Subway entrance\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001806", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Urban\", \"C\": \"Subway entrance\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001806", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Strong shadows\", \"C\": \"Flashing lights\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001807", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Strong shadows\", \"C\": \"Flashing lights\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001807", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001808", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001808", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Clear\", \"C\": \"Strong wind\", \"D\": \"Snowy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001809", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Clear\", \"C\": \"Strong wind\", \"D\": \"Snowy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001809", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Countdown timer visible\", \"C\": \"Blinking green\", \"D\": \"Flashing amber\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001810", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Countdown timer visible\", \"C\": \"Blinking green\", \"D\": \"Flashing amber\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001810", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Wet\", \"C\": \"Potholes\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001811", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/666.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Wet\", \"C\": \"Potholes\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001811", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Coastal road\", \"C\": \"Bridge\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001812", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Coastal road\", \"C\": \"Bridge\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001812", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Clear\", \"C\": \"Backlit\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001813", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dawn\", \"B\": \"Clear\", \"C\": \"Backlit\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001813", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like on the road? You should answer one choice from four answers \"A\": \"Empty\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001814", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like on the road? You should answer one choice from four answers \"A\": \"Empty\", \"B\": \"Congested\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Empty" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001814", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Flurries\", \"C\": \"Cloudy\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001815", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Smoggy\", \"B\": \"Flurries\", \"C\": \"Cloudy\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001815", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Yellow\", \"C\": \"Blinking green\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001816", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Yellow\", \"C\": \"Blinking green\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001816", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Dry\", \"C\": \"Muddy\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001817", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/675.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Water puddles\", \"B\": \"Dry\", \"C\": \"Muddy\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001817", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Construction zone\", \"C\": \"Roundabout\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001818", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Construction zone\", \"C\": \"Roundabout\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001818", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low\", \"C\": \"Dusk\", \"D\": \"Partly cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001819", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low\", \"C\": \"Dusk\", \"D\": \"Partly cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001819", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like in this scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001820", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like in this scene? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Accident blockage\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001820", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Strong wind\", \"C\": \"Clear\", \"D\": \"Cloudy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001821", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Strong wind\", \"C\": \"Clear\", \"D\": \"Cloudy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001821", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Unknown\", \"C\": \"Green\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001822", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Unknown\", \"C\": \"Green\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001822", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Dry\", \"C\": \"Leaf-covered\", \"D\": \"Gravel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001823", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/684.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Uneven\", \"B\": \"Dry\", \"C\": \"Leaf-covered\", \"D\": \"Gravel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001823", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Urban\", \"C\": \"Campus area\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001824", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Harbor area\", \"B\": \"Urban\", \"C\": \"Campus area\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001824", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Warm\", \"B\": \"Rain-blurred\", \"C\": \"Glare\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001825", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in the scene? You should answer one choice from four answers \"A\": \"Warm\", \"B\": \"Rain-blurred\", \"C\": \"Glare\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001825", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like in this scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001826", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic flow like in this scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001826", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Dry\", \"C\": \"Snowy\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001827", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Mist\", \"B\": \"Dry\", \"C\": \"Snowy\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001827", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Off / malfunctioning\", \"C\": \"Not visible\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001828", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Off / malfunctioning\", \"C\": \"Not visible\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001828", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described in the scene? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Water puddles\", \"C\": \"Smooth\", \"D\": \"Newly paved\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001829", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/693.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described in the scene? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Water puddles\", \"C\": \"Smooth\", \"D\": \"Newly paved\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001829", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Airport roadway\", \"C\": \"Roundabout\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001830", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is depicted in the scene? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Airport roadway\", \"C\": \"Roundabout\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001830", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Glare\", \"C\": \"Strong shadows\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001831", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is visibility in this nighttime scene? You should answer one choice from four answers \"A\": \"Poor\", \"B\": \"Glare\", \"C\": \"Strong shadows\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001831", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Congested\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001832", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Accident blockage\", \"C\": \"Congested\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001832", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Sunny\", \"C\": \"Drizzling\", \"D\": \"Rain\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001833", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Strong wind\", \"B\": \"Sunny\", \"C\": \"Drizzling\", \"D\": \"Rain\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rain" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001833", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Red\", \"C\": \"Countdown timer visible\", \"D\": \"Unknown\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001834", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Red\", \"C\": \"Countdown timer visible\", \"D\": \"Unknown\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001834", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Icy\", \"C\": \"Potholes\", \"D\": \"Flooded\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001835", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/701.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Wet\", \"B\": \"Icy\", \"C\": \"Potholes\", \"D\": \"Flooded\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Wet" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001835", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Construction zone\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001836", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Construction zone\", \"C\": \"Urban\", \"D\": \"Service station\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001836", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001837", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Foggy\", \"C\": \"Clear\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001837", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001838", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Moderate" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001838", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Dry\", \"C\": \"Strong wind\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001839", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Dry\", \"C\": \"Strong wind\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001839", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"None\", \"C\": \"Yellow\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001840", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the frame? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"None\", \"C\": \"Yellow\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001840", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Dry\", \"C\": \"Icy\", \"D\": \"Bumpy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001841", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/710.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Sandy\", \"B\": \"Dry\", \"C\": \"Icy\", \"D\": \"Bumpy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001841", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Airport roadway\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001842", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Airport roadway\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001842", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Strong shadows\", \"C\": \"Clear\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001843", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Strong shadows\", \"C\": \"Clear\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001843", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001844", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Moderate\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001844", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rainy\", \"C\": \"Mist\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001845", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rainy\", \"C\": \"Mist\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001845", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Not visible\", \"C\": \"Pedestrian signal\", \"D\": \"Dim / unclear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001846", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in the image? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"Not visible\", \"C\": \"Pedestrian signal\", \"D\": \"Dim / unclear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001846", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Potholes\", \"C\": \"Smooth\", \"D\": \"Uneven\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001847", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/72.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Potholes\", \"C\": \"Smooth\", \"D\": \"Uneven\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001847", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Two-lane street\", \"B\": \"Highway\", \"C\": \"Parking lot\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001848", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Two-lane street\", \"B\": \"Highway\", \"C\": \"Parking lot\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Two-lane street" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001848", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Low\", \"C\": \"Reflected glare\", \"D\": \"Flashing lights\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001849", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Strong shadows\", \"B\": \"Low\", \"C\": \"Reflected glare\", \"D\": \"Flashing lights\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001849", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001850", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001850", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Windy\", \"C\": \"Overcast\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001851", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Stormy\", \"B\": \"Windy\", \"C\": \"Overcast\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Overcast" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001851", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Covered by tree\", \"C\": \"Countdown timer visible\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001852", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible or active? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Covered by tree\", \"C\": \"Countdown timer visible\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001852", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Snowy\", \"C\": \"Snow-covered\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001853", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/729.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Partially damaged\", \"B\": \"Snowy\", \"C\": \"Snow-covered\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001853", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001854", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Highway\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001854", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is lighting affecting the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Low light\", \"C\": \"Poor\", \"D\": \"Overexposed\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001855", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is lighting affecting the scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Low light\", \"C\": \"Poor\", \"D\": \"Overexposed\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Poor" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001855", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Two\", \"C\": \"Moderate\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001856", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How many vehicles are visible in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Two\", \"C\": \"Moderate\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Two" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001856", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any indication of weather conditions? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dry\", \"C\": \"Gloomy\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001857", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any indication of weather conditions? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Dry\", \"C\": \"Gloomy\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001857", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Blinking green\", \"C\": \"Red\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001858", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Can the traffic light status be determined from the scene? You should answer one choice from four answers \"A\": \"Unknown\", \"B\": \"Blinking green\", \"C\": \"Red\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Unknown" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001858", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Newly paved\", \"C\": \"Smooth\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001859", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/738.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Construction debris\", \"B\": \"Newly paved\", \"C\": \"Smooth\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001859", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Airport roadway\", \"C\": \"Subway entrance\", \"D\": \"Bridge\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001860", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Airport roadway\", \"C\": \"Subway entrance\", \"D\": \"Bridge\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001860", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Strong shadows\", \"C\": \"Partly cloudy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001861", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Blinding sunlight\", \"B\": \"Strong shadows\", \"C\": \"Partly cloudy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001861", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001862", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Congested\", \"C\": \"Accident blockage\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001862", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Sleet\", \"C\": \"Gloomy\", \"D\": \"Sunny\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001863", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Sleet\", \"C\": \"Gloomy\", \"D\": \"Sunny\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001863", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing yellow\", \"C\": \"Countdown timer visible\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001864", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Not visible\", \"B\": \"Flashing yellow\", \"C\": \"Countdown timer visible\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001864", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Newly paved\", \"C\": \"Dry\", \"D\": \"Muddy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001865", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/747.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Potholes\", \"B\": \"Newly paved\", \"C\": \"Dry\", \"D\": \"Muddy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001865", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Urban\", \"C\": \"Forest road\", \"D\": \"Subway entrance\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001866", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this driving scene set in? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Urban\", \"C\": \"Forest road\", \"D\": \"Subway entrance\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001866", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low light\", \"C\": \"Under tunnel\", \"D\": \"Backlit\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001867", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Low light\", \"C\": \"Under tunnel\", \"D\": \"Backlit\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001867", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in this scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Accident blockage\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001868", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity in this scene? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Accident blockage\", \"C\": \"Congested\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001868", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Cloudy\", \"C\": \"Clear\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001869", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Drizzling\", \"B\": \"Cloudy\", \"C\": \"Clear\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001869", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Reflected signal\", \"C\": \"Flashing red\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001870", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Reflected signal\", \"C\": \"Flashing red\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001870", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Dry\", \"C\": \"Blocked by obstacle\", \"D\": \"Debris scattered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001871", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/756.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Oil slick\", \"B\": \"Dry\", \"C\": \"Blocked by obstacle\", \"D\": \"Debris scattered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001871", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Urban\", \"C\": \"Suburban\", \"D\": \"Airport roadway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001872", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Bridge\", \"B\": \"Urban\", \"C\": \"Suburban\", \"D\": \"Airport roadway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Suburban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001872", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hazy\", \"C\": \"Dusk\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001873", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Hazy\", \"C\": \"Dusk\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001873", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001874", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Rush hour\", \"C\": \"Moderate\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001874", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Strong wind\", \"C\": \"Windy\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001875", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Strong wind\", \"C\": \"Windy\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001875", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Flashing red\", \"C\": \"None\", \"D\": \"Hidden\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001876", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing amber\", \"B\": \"Flashing red\", \"C\": \"None\", \"D\": \"Hidden\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hidden" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001876", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Bumpy\", \"C\": \"Dry\", \"D\": \"Gravel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001877", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/765.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Bumpy\", \"C\": \"Dry\", \"D\": \"Gravel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001877", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Service station\", \"C\": \"Two-lane road\", \"D\": \"Forest road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001878", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Rural\", \"B\": \"Service station\", \"C\": \"Two-lane road\", \"D\": \"Forest road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Two-lane road" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001878", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Excellent\", \"B\": \"Dim streetlight visibility\", \"C\": \"Glare\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001879", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Excellent\", \"B\": \"Dim streetlight visibility\", \"C\": \"Glare\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Excellent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001879", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001880", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic condition? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Heavy\", \"C\": \"Light\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001880", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Windy\", \"C\": \"Dusty\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001881", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Windy\", \"C\": \"Dusty\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001881", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001882", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic signals visible in the scene? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"None\", \"C\": \"Reflected signal\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001882", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rough surface\", \"C\": \"Bumpy\", \"D\": \"Uneven\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001883", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/774.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Rough surface\", \"C\": \"Bumpy\", \"D\": \"Uneven\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001883", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Suburban\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001884", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Suburban\", \"C\": \"Parking lot\", \"D\": \"Urban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001884", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Partly cloudy\", \"C\": \"Clear\", \"D\": \"Low light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001885", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Partly cloudy\", \"C\": \"Clear\", \"D\": \"Low light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Low light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001885", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001886", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Lane merging\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001886", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Overcast\", \"C\": \"Mist\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001887", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Overcast\", \"C\": \"Mist\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001887", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Green\", \"C\": \"Flashing red\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001888", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Green\", \"C\": \"Flashing red\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001888", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Potholes\", \"C\": \"Newly paved\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001889", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/783.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Potholes\", \"C\": \"Newly paved\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001889", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Service station\", \"C\": \"Campus area\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001890", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road setting is depicted? You should answer one choice from four answers \"A\": \"Coastal road\", \"B\": \"Service station\", \"C\": \"Campus area\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001890", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Hazy\", \"C\": \"Overexposed\", \"D\": \"Excellent\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001891", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Low light\", \"B\": \"Hazy\", \"C\": \"Overexposed\", \"D\": \"Excellent\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Excellent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001891", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001892", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Detour traffic\", \"C\": \"Light\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001892", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Windy\", \"C\": \"Sunny\", \"D\": \"Mist\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001893", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Foggy\", \"B\": \"Windy\", \"C\": \"Sunny\", \"D\": \"Mist\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001893", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Reflected signal\", \"C\": \"None\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001894", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Flashing yellow\", \"B\": \"Reflected signal\", \"C\": \"None\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001894", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Dry\", \"C\": \"Construction debris\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001895", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/792.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Icy\", \"B\": \"Dry\", \"C\": \"Construction debris\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001895", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Forest road\", \"C\": \"Construction zone\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001896", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of road is depicted in the scene? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Forest road\", \"C\": \"Construction zone\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001896", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Excellent\", \"C\": \"Dim streetlight visibility\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001897", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Excellent\", \"C\": \"Dim streetlight visibility\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Excellent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001897", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any visible traffic in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"None\", \"C\": \"Detour traffic\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001898", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any visible traffic in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"None\", \"C\": \"Detour traffic\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001898", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Rainy\", \"C\": \"Thunderstorm\", \"D\": \"Dry\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001899", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Rainy\", \"C\": \"Thunderstorm\", \"D\": \"Dry\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001899", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Blinking green\", \"C\": \"Flashing yellow\", \"D\": \"No\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001900", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Blinking green\", \"C\": \"Flashing yellow\", \"D\": \"No\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001900", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Dry\", \"C\": \"Snowy\", \"D\": \"Blocked by obstacle\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001901", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/800.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How does the road surface appear? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Dry\", \"C\": \"Snowy\", \"D\": \"Blocked by obstacle\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001901", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Residential\", \"C\": \"Shopping district\", \"D\": \"Suburban\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001902", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Residential\", \"C\": \"Shopping district\", \"D\": \"Suburban\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001902", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Clear\", \"C\": \"Dawn\", \"D\": \"Smoky\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001903", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Clear\", \"C\": \"Dawn\", \"D\": \"Smoky\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001903", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there traffic on the road? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Accident blockage\", \"C\": \"None\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001904", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there traffic on the road? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Accident blockage\", \"C\": \"None\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001904", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather like? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Sunny\", \"C\": \"Sleet\", \"D\": \"Hailing\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001905", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather like? You should answer one choice from four answers \"A\": \"Sandstorm\", \"B\": \"Sunny\", \"C\": \"Sleet\", \"D\": \"Hailing\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001905", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Green\", \"C\": \"Flashing yellow\", \"D\": \"Countdown timer visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001906", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Green\", \"C\": \"Flashing yellow\", \"D\": \"Countdown timer visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001906", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Gravel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001907", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/81.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Gravel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001907", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Residential\", \"C\": \"Campus area\", \"D\": \"Shopping district\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001908", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Forest road\", \"B\": \"Residential\", \"C\": \"Campus area\", \"D\": \"Shopping district\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001908", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Overexposed\", \"C\": \"Under tunnel\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001909", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Dusk\", \"B\": \"Overexposed\", \"C\": \"Under tunnel\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Overexposed" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001909", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001910", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Congested\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001910", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Sunny\", \"C\": \"Sleet\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001911", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Thunderstorm\", \"B\": \"Sunny\", \"C\": \"Sleet\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001911", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Absent\", \"C\": \"Countdown timer visible\", \"D\": \"Blinking green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001912", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Yellow\", \"B\": \"Absent\", \"C\": \"Countdown timer visible\", \"D\": \"Blinking green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001912", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Cracked asphalt\", \"C\": \"Dry\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001913", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/819.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Cracked asphalt\", \"C\": \"Dry\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001913", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Coastal road\", \"C\": \"Roundabout\", \"D\": \"Construction zone\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001914", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Residential\", \"B\": \"Coastal road\", \"C\": \"Roundabout\", \"D\": \"Construction zone\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001914", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Backlit\", \"C\": \"Dim\", \"D\": \"Dawn\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001915", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Backlit\", \"C\": \"Dim\", \"D\": \"Dawn\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dim" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001915", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001916", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of traffic activity? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001916", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Sunny\", \"C\": \"Thunderstorm\", \"D\": \"Gloomy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001917", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Sunny\", \"C\": \"Thunderstorm\", \"D\": \"Gloomy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001917", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"No\", \"C\": \"Red\", \"D\": \"Reflected signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001918", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"No\", \"C\": \"Red\", \"D\": \"Reflected signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001918", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Blocked by obstacle\", \"C\": \"Rough surface\", \"D\": \"Cracked asphalt\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001919", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/828.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Blocked by obstacle\", \"C\": \"Rough surface\", \"D\": \"Cracked asphalt\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001919", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Urban\", \"C\": \"Campus area\", \"D\": \"Market street\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001920", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Urban\", \"C\": \"Campus area\", \"D\": \"Market street\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001920", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Reflected glare\", \"C\": \"Dawn\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001921", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Glare\", \"B\": \"Reflected glare\", \"C\": \"Dawn\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001921", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001922", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density? You should answer one choice from four answers \"A\": \"Heavy\", \"B\": \"Light\", \"C\": \"Moderate\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001922", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Drizzling\", \"C\": \"Dry\", \"D\": \"Smoggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001923", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Snowy\", \"B\": \"Drizzling\", \"C\": \"Dry\", \"D\": \"Smoggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001923", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic signal visible? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Countdown timer visible\", \"C\": \"Not visible\", \"D\": \"Covered by tree\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001924", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic signal visible? You should answer one choice from four answers \"A\": \"Temporary road sign light\", \"B\": \"Countdown timer visible\", \"C\": \"Not visible\", \"D\": \"Covered by tree\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001924", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Bumpy\", \"C\": \"Smooth\", \"D\": \"Debris scattered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001925", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/837.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the surface condition of the road? You should answer one choice from four answers \"A\": \"Newly paved\", \"B\": \"Bumpy\", \"C\": \"Smooth\", \"D\": \"Debris scattered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001925", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001926", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this road in? You should answer one choice from four answers \"A\": \"Roundabout\", \"B\": \"Coastal road\", \"C\": \"Urban\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001926", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Glare\", \"C\": \"Reflected glare\", \"D\": \"Sepia\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001927", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Under tunnel\", \"B\": \"Glare\", \"C\": \"Reflected glare\", \"D\": \"Sepia\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sepia" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001927", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001928", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like in this scene? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Moderate\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001928", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snowstorm\", \"C\": \"Hailing\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001929", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Snowstorm\", \"C\": \"Hailing\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001929", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Not visible\", \"C\": \"Reflected signal\", \"D\": \"Temporary road sign light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001930", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Red\", \"B\": \"Not visible\", \"C\": \"Reflected signal\", \"D\": \"Temporary road sign light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001930", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Dry\", \"C\": \"Icy\", \"D\": \"Water puddles\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001931", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/846.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Dry\", \"C\": \"Icy\", \"D\": \"Water puddles\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001931", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Highway\", \"C\": \"Rural\", \"D\": \"Residential\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001932", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this street located in? You should answer one choice from four answers \"A\": \"Campus area\", \"B\": \"Highway\", \"C\": \"Rural\", \"D\": \"Residential\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001932", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Low light\", \"C\": \"Glare\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001933", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Hazy\", \"B\": \"Low light\", \"C\": \"Glare\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001933", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Rush hour\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001934", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this scene? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Light\", \"C\": \"Congested\", \"D\": \"Rush hour\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001934", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Foggy\", \"C\": \"Dusty\", \"D\": \"Rainy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001935", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present in the scene? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Foggy\", \"C\": \"Dusty\", \"D\": \"Rainy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001935", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Missing\", \"C\": \"Blinking green\", \"D\": \"Off / malfunctioning\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001936", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is the traffic light visible in this scene? You should answer one choice from four answers \"A\": \"Flashing red\", \"B\": \"Missing\", \"C\": \"Blinking green\", \"D\": \"Off / malfunctioning\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Missing" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001936", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Gravel\", \"C\": \"Potholes\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001937", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/855.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Cracked asphalt\", \"B\": \"Gravel\", \"C\": \"Potholes\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001937", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Campus area\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001938", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Shopping district\", \"B\": \"Campus area\", \"C\": \"Urban\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001938", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting condition? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Reflected glare\", \"C\": \"Overexposed\", \"D\": \"Under tunnel\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001939", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting condition? You should answer one choice from four answers \"A\": \"Smoky\", \"B\": \"Reflected glare\", \"C\": \"Overexposed\", \"D\": \"Under tunnel\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Overexposed" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001939", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001940", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic density like? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Rush hour\", \"C\": \"Accident blockage\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001940", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Thunderstorm\", \"C\": \"Overcast\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001941", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Thunderstorm\", \"C\": \"Overcast\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001941", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Countdown timer visible\", \"C\": \"Blinking green\", \"D\": \"Pedestrian signal\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001942", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Countdown timer visible\", \"C\": \"Blinking green\", \"D\": \"Pedestrian signal\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001942", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001943", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/864.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the road surface like? You should answer one choice from four answers \"A\": \"Bumpy\", \"B\": \"Water puddles\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001943", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Harbor area\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001944", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of environment is this? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Urban\", \"C\": \"Parking lot\", \"D\": \"Harbor area\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001944", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Reflected glare\", \"C\": \"Hazy\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001945", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Reflected glare\", \"C\": \"Hazy\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001945", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001946", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the traffic condition like? You should answer one choice from four answers \"A\": \"Lane merging\", \"B\": \"Moderate\", \"C\": \"Rush hour\", \"D\": \"Light\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001946", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Dusty\", \"C\": \"Mist\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001947", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Rainy\", \"B\": \"Dusty\", \"C\": \"Mist\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001947", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Flashing yellow\", \"C\": \"Yellow\", \"D\": \"Not visible\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001948", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Dim / unclear\", \"B\": \"Flashing yellow\", \"C\": \"Yellow\", \"D\": \"Not visible\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Not visible" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001948", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Wet\", \"C\": \"Leaf-covered\", \"D\": \"Uneven\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001949", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/873.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Smooth\", \"B\": \"Wet\", \"C\": \"Leaf-covered\", \"D\": \"Uneven\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001949", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Construction zone\", \"C\": \"Highway\", \"D\": \"Coastal road\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001950", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Urban\", \"B\": \"Construction zone\", \"C\": \"Highway\", \"D\": \"Coastal road\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001950", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Dawn\", \"C\": \"Hazy\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001951", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting and clarity in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Dawn\", \"C\": \"Hazy\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Hazy" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001951", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this area? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Lane merging\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001952", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the level of vehicle movement in this area? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"Detour traffic\", \"C\": \"Rush hour\", \"D\": \"Lane merging\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001952", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Overcast\", \"C\": \"Drizzling\", \"D\": \"Clear\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001953", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there any weather affecting the scene? You should answer one choice from four answers \"A\": \"Cloudy\", \"B\": \"Overcast\", \"C\": \"Drizzling\", \"D\": \"Clear\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001953", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Flashing amber\", \"C\": \"Green\", \"D\": \"Absent\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001954", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in this scene? You should answer one choice from four answers \"A\": \"Countdown timer visible\", \"B\": \"Flashing amber\", \"C\": \"Green\", \"D\": \"Absent\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Absent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001954", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Flooded\", \"C\": \"Muddy\", \"D\": \"Slippery\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001955", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/882.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Flooded\", \"C\": \"Muddy\", \"D\": \"Slippery\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001955", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Residential\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001956", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Subway entrance\", \"B\": \"Residential\", \"C\": \"Urban\", \"D\": \"Highway\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Residential" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001956", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Blinding sunlight\", \"C\": \"Clear\", \"D\": \"Hazy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001957", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Backlit\", \"B\": \"Blinding sunlight\", \"C\": \"Clear\", \"D\": \"Hazy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001957", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001958", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How busy is the traffic on this road? You should answer one choice from four answers \"A\": \"Rush hour\", \"B\": \"Lane merging\", \"C\": \"Light\", \"D\": \"Heavy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001958", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Overcast\", \"C\": \"Drizzling\", \"D\": \"Foggy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001959", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather conditions are present? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Overcast\", \"C\": \"Drizzling\", \"D\": \"Foggy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001959", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Dim / unclear\", \"C\": \"Temporary road sign light\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001960", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are there any visible traffic signals? You should answer one choice from four answers \"A\": \"None\", \"B\": \"Dim / unclear\", \"C\": \"Temporary road sign light\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001960", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Uneven\", \"C\": \"Slippery\", \"D\": \"Smooth\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001961", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/891.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Snow-covered\", \"B\": \"Uneven\", \"C\": \"Slippery\", \"D\": \"Smooth\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Smooth" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001961", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Airport roadway\", \"C\": \"Market street\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001962", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Parking lot\", \"B\": \"Airport roadway\", \"C\": \"Market street\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001962", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Clear\", \"C\": \"Hazy\", \"D\": \"Rain-blurred\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001963", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Dim streetlight visibility\", \"B\": \"Clear\", \"C\": \"Hazy\", \"D\": \"Rain-blurred\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Clear" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001963", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic situation? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001964", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic situation? You should answer one choice from four answers \"A\": \"Moderate\", \"B\": \"Rush hour\", \"C\": \"Light\", \"D\": \"Detour traffic\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001964", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Strong wind\", \"C\": \"Mist\", \"D\": \"Sandstorm\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001965", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Strong wind\", \"C\": \"Mist\", \"D\": \"Sandstorm\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001965", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"No\", \"C\": \"Temporary road sign light\", \"D\": \"Flashing red\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001966", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Covered by tree\", \"B\": \"No\", \"C\": \"Temporary road sign light\", \"D\": \"Flashing red\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001966", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Uneven\", \"C\": \"Leaf-covered\", \"D\": \"Construction debris\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001967", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/90.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Uneven\", \"C\": \"Leaf-covered\", \"D\": \"Construction debris\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001967", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Roundabout\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001968", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is this road located in? You should answer one choice from four answers \"A\": \"Market street\", \"B\": \"Bridge\", \"C\": \"Urban\", \"D\": \"Roundabout\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Urban" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001968", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Clear\", \"C\": \"Rain-blurred\", \"D\": \"Warm haze\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001969", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the lighting in this scene? You should answer one choice from four answers \"A\": \"Overexposed\", \"B\": \"Clear\", \"C\": \"Rain-blurred\", \"D\": \"Warm haze\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Warm haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001969", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001970", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current traffic density like? You should answer one choice from four answers \"A\": \"Detour traffic\", \"B\": \"Light\", \"C\": \"Rush hour\", \"D\": \"Accident blockage\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Light" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001970", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Sandstorm\", \"C\": \"Windy\", \"D\": \"Flurries\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001971", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What weather condition is present? You should answer one choice from four answers \"A\": \"Haze\", \"B\": \"Sandstorm\", \"C\": \"Windy\", \"D\": \"Flurries\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Haze" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001971", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Red\", \"C\": \"Reflected signal\", \"D\": \"No\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001972", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the scene? You should answer one choice from four answers \"A\": \"Off / malfunctioning\", \"B\": \"Red\", \"C\": \"Reflected signal\", \"D\": \"No\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001972", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Cracked asphalt\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "C" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001973", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/909.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the road surface described? You should answer one choice from four answers \"A\": \"Muddy\", \"B\": \"Cracked asphalt\", \"C\": \"Dry\", \"D\": \"Snow-covered\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001973", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Parking lot\", \"C\": \"Roundabout\", \"D\": \"Rural\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001974", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Scene type", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What kind of area is depicted in the scene? You should answer one choice from four answers \"A\": \"Construction zone\", \"B\": \"Parking lot\", \"C\": \"Roundabout\", \"D\": \"Rural\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Rural" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001974", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Strong shadows\", \"D\": \"Excellent\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "D" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001975", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Visibility", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. How is the visibility in this scene? You should answer one choice from four answers \"A\": \"Clear\", \"B\": \"Foggy\", \"C\": \"Strong shadows\", \"D\": \"Excellent\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Excellent" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001975", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there traffic present in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"None\", \"C\": \"Accident blockage\", \"D\": \"Congested\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "B" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001976", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Traffic flow", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Is there traffic present in the scene? You should answer one choice from four answers \"A\": \"Light\", \"B\": \"None\", \"C\": \"Accident blockage\", \"D\": \"Congested\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "None" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001976", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Foggy\", \"C\": \"Strong wind\", \"D\": \"Windy\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001977", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Weather", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the current weather condition? You should answer one choice from four answers \"A\": \"Sunny\", \"B\": \"Foggy\", \"C\": \"Strong wind\", \"D\": \"Windy\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Sunny" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001977", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Blinking green\", \"C\": \"Yellow\", \"D\": \"Green\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001978", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Traffic light", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. Are traffic lights visible in the frame? You should answer one choice from four answers \"A\": \"No\", \"B\": \"Blinking green\", \"C\": \"Yellow\", \"D\": \"Green\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "No" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001978", "subtask": "label_text" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Cracked asphalt\", \"C\": \"Gravel\", \"D\": \"Partially damaged\". Respond with ONLY the letter of the selected option: A, B, C, or D. Do not include the option text, punctuation, or explanation. For example, respond exactly as 'A'." }, { "from": "gpt", "value": "A" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001979", "subtask": "option_letter" }, { "image": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.png", "event": "data/perception/pku/aps_frames_sampled/val/motion_blur/010_val_motion_blur/93.npz", "category": "Road condition", "conversations": [ { "from": "human", "value": "Suppose you are driving, and I'm providing you with the image captured by the car's front center. What is the state of the road surface? You should answer one choice from four answers \"A\": \"Dry\", \"B\": \"Cracked asphalt\", \"C\": \"Gravel\", \"D\": \"Partially damaged\". Respond with ONLY the corresponding option text exactly as written. Do not include the option letter or any explanation. For example, respond exactly as 'Urban'." }, { "from": "gpt", "value": "Dry" } ], "original_id": "perception/pku/pku_test_perception_hard.json:001979", "subtask": "label_text" } ]