Dataset Viewer
Auto-converted to Parquet
images
images listlengths
1
1
data_source
stringclasses
1 value
reward_model
dict
extra_info
dict
prompt
listlengths
2
2
ability
stringlengths
20
31
env_name
stringclasses
1 value
infovqa_augmented
{ "ground_truth": "Tony Blair." }
{ "all_answers": [ "Tony Blair." ], "answer": "Tony Blair.", "augmentation_type": "rot270+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_0e373ef693ab6534", "index": "infovqa_infovqa_000000_rot270_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "0e373ef693ab65342b07a76926d4b3ce475306b37679d774f79ab4fc25619fa7", "original_question_id": "infovqa_000000", "question": "What is the name of the famous person who was thrown with condoms?\nKeep it brief.", "question_id": "infovqa_000000" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the name of the famous person who was thrown with condoms?\nKeep it brief.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot270_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "1964." }
{ "all_answers": [ "1964." ], "answer": "1964.", "augmentation_type": "rot90", "data_split": "train", "doc_id": "Infographic-VQA_24e22b1bfebd454f", "index": "infovqa_infovqa_000001_rot90", "infovqa_source": "Infographic-VQA", "original_image_filename": "24e22b1bfebd454f9d95400da2bd86f73f7fd9d430dcb0a869d0fa6e2e061018", "original_question_id": "infovqa_000001", "question": "in which year was the skateboard the longest\nAnswer briefly.", "question_id": "infovqa_000001" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nin which year was the skateboard the longest\nAnswer briefly.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_temporal_rot90
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Road rider crusier." }
{ "all_answers": [ "Road rider crusier." ], "answer": "Road rider crusier.", "augmentation_type": "rot90+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_24e22b1bfebd454f", "index": "infovqa_infovqa_000002_rot90_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "24e22b1bfebd454f9d95400da2bd86f73f7fd9d430dcb0a869d0fa6e2e061018", "original_question_id": "infovqa_000002", "question": "what was the name of the skateboard with precision bearings\nQuick response, please.", "question_id": "infovqa_000002" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwhat was the name of the skateboard with precision bearings\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot90_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "10." }
{ "all_answers": [ "10." ], "answer": "10.", "augmentation_type": "flip_v", "data_split": "train", "doc_id": "Infographic-VQA_24e22b1bfebd454f", "index": "infovqa_infovqa_000003_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "24e22b1bfebd454f9d95400da2bd86f73f7fd9d430dcb0a869d0fa6e2e061018", "original_question_id": "infovqa_000003", "question": "How many years did it take after 1990 for the new design\nShort answer required.", "question_id": "infovqa_000003" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many years did it take after 1990 for the new design\nShort answer required.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_temporal_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "G and S Fiberflex Pintail." }
{ "all_answers": [ "G and S Fiberflex Pintail." ], "answer": "G and S Fiberflex Pintail.", "augmentation_type": "rot180", "data_split": "train", "doc_id": "Infographic-VQA_24e22b1bfebd454f", "index": "infovqa_infovqa_000004_rot180", "infovqa_source": "Infographic-VQA", "original_image_filename": "24e22b1bfebd454f9d95400da2bd86f73f7fd9d430dcb0a869d0fa6e2e061018", "original_question_id": "infovqa_000004", "question": "What was the name of the board introduced in 1964\nConcise answer only.", "question_id": "infovqa_000004" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat was the name of the board introduced in 1964\nConcise answer only.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "White." }
{ "all_answers": [ "White." ], "answer": "White.", "augmentation_type": "rot90+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_24e22b1bfebd454f", "index": "infovqa_infovqa_000005_rot90_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "24e22b1bfebd454f9d95400da2bd86f73f7fd9d430dcb0a869d0fa6e2e061018", "original_question_id": "infovqa_000005", "question": "what is the colour of all the boards, white or blue?\nQuick response, please.", "question_id": "infovqa_000005" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwhat is the colour of all the boards, white or blue?\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot90_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "1980." }
{ "all_answers": [ "1980." ], "answer": "1980.", "augmentation_type": "rot180+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_24e22b1bfebd454f", "index": "infovqa_infovqa_000006_rot180_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "24e22b1bfebd454f9d95400da2bd86f73f7fd9d430dcb0a869d0fa6e2e061018", "original_question_id": "infovqa_000006", "question": "when was the old school fishtail introduced?\nKeep it brief.", "question_id": "infovqa_000006" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwhen was the old school fishtail introduced?\nKeep it brief.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_temporal_rot180_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "1960." }
{ "all_answers": [ "1960." ], "answer": "1960.", "augmentation_type": "rot180", "data_split": "train", "doc_id": "Infographic-VQA_24e22b1bfebd454f", "index": "infovqa_infovqa_000007_rot180", "infovqa_source": "Infographic-VQA", "original_image_filename": "24e22b1bfebd454f9d95400da2bd86f73f7fd9d430dcb0a869d0fa6e2e061018", "original_question_id": "infovqa_000007", "question": "when was nash shark introduced?\nAnswer briefly.", "question_id": "infovqa_000007" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwhen was nash shark introduced?\nAnswer briefly.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_temporal_rot180
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "12." }
{ "all_answers": [ "12." ], "answer": "12.", "augmentation_type": "rot180+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_84ddf5ccb6264b23", "index": "infovqa_infovqa_000008_rot180_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "84ddf5ccb6264b23324ef25ad0131dcbcad80fda3c9d67ab4a4098155a8505b8", "original_question_id": "infovqa_000008", "question": "how many enemies does the doctor have\nShort answer required.", "question_id": "infovqa_000008" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nhow many enemies does the doctor have\nShort answer required.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot180_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "The skarasen." }
{ "all_answers": [ "The skarasen." ], "answer": "The skarasen.", "augmentation_type": "flip_h", "data_split": "train", "doc_id": "Infographic-VQA_84ddf5ccb6264b23", "index": "infovqa_infovqa_000009_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "84ddf5ccb6264b23324ef25ad0131dcbcad80fda3c9d67ab4a4098155a8505b8", "original_question_id": "infovqa_000009", "question": "who is often mistaken for the lock ness monster\nConcise answer only.", "question_id": "infovqa_000009" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwho is often mistaken for the lock ness monster\nConcise answer only.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Ice warriors." }
{ "all_answers": [ "Ice warriors." ], "answer": "Ice warriors.", "augmentation_type": "flip_v", "data_split": "train", "doc_id": "Infographic-VQA_84ddf5ccb6264b23", "index": "infovqa_infovqa_000010_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "84ddf5ccb6264b23324ef25ad0131dcbcad80fda3c9d67ab4a4098155a8505b8", "original_question_id": "infovqa_000010", "question": "who are extremely sensitive to heat\nConcise answer only.", "question_id": "infovqa_000010" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwho are extremely sensitive to heat\nConcise answer only.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "10." }
{ "all_answers": [ "10." ], "answer": "10.", "augmentation_type": "rot270+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_84ddf5ccb6264b23", "index": "infovqa_infovqa_000011_rot270_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "84ddf5ccb6264b23324ef25ad0131dcbcad80fda3c9d67ab4a4098155a8505b8", "original_question_id": "infovqa_000011", "question": "the enemies identified are in a period of how many years\nKeep it brief.", "question_id": "infovqa_000011" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nthe enemies identified are in a period of how many years\nKeep it brief.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_temporal_rot270_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Blue." }
{ "all_answers": [ "Blue." ], "answer": "Blue.", "augmentation_type": "rot270", "data_split": "train", "doc_id": "Infographic-VQA_84ddf5ccb6264b23", "index": "infovqa_infovqa_000012_rot270", "infovqa_source": "Infographic-VQA", "original_image_filename": "84ddf5ccb6264b23324ef25ad0131dcbcad80fda3c9d67ab4a4098155a8505b8", "original_question_id": "infovqa_000012", "question": "what is the colour of the police box, blue, red?\nBe succinct.", "question_id": "infovqa_000012" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwhat is the colour of the police box, blue, red?\nBe succinct.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot270
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Sontarans." }
{ "all_answers": [ "Sontarans." ], "answer": "Sontarans.", "augmentation_type": "rot180", "data_split": "train", "doc_id": "Infographic-VQA_84ddf5ccb6264b23", "index": "infovqa_infovqa_000013_rot180", "infovqa_source": "Infographic-VQA", "original_image_filename": "84ddf5ccb6264b23324ef25ad0131dcbcad80fda3c9d67ab4a4098155a8505b8", "original_question_id": "infovqa_000013", "question": "who is holding a helmet in his hand?\nQuick response, please.", "question_id": "infovqa_000013" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwho is holding a helmet in his hand?\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Tim Hortons." }
{ "all_answers": [ "Tim Hortons." ], "answer": "Tim Hortons.", "augmentation_type": "rot180+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_98fe47f5c7b5c509", "index": "infovqa_infovqa_000014_rot180_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "98fe47f5c7b5c50984efca4babeb2a224cadaab79ebcfae743a22b23c5aeb3a1", "original_question_id": "infovqa_000014", "question": "where should I eat if I am in Canada\nShort answer required.", "question_id": "infovqa_000014" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwhere should I eat if I am in Canada\nShort answer required.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_spatial_rot180_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Taco bell." }
{ "all_answers": [ "Taco bell." ], "answer": "Taco bell.", "augmentation_type": "rot90", "data_split": "train", "doc_id": "Infographic-VQA_98fe47f5c7b5c509", "index": "infovqa_infovqa_000015_rot90", "infovqa_source": "Infographic-VQA", "original_image_filename": "98fe47f5c7b5c50984efca4babeb2a224cadaab79ebcfae743a22b23c5aeb3a1", "original_question_id": "infovqa_000015", "question": "where should I eat if I dont have more than $3\nAnswer briefly.", "question_id": "infovqa_000015" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwhere should I eat if I dont have more than $3\nAnswer briefly.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_spatial_rot90
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Burger King." }
{ "all_answers": [ "Burger King." ], "answer": "Burger King.", "augmentation_type": "rot270+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_98fe47f5c7b5c509", "index": "infovqa_infovqa_000016_rot270_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "98fe47f5c7b5c50984efca4babeb2a224cadaab79ebcfae743a22b23c5aeb3a1", "original_question_id": "infovqa_000016", "question": "where should I eat if I am not older than 16?\nKeep it brief.", "question_id": "infovqa_000016" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwhere should I eat if I am not older than 16?\nKeep it brief.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_spatial_rot270_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "McDonalds." }
{ "all_answers": [ "McDonalds." ], "answer": "McDonalds.", "augmentation_type": "flip_h", "data_split": "train", "doc_id": "Infographic-VQA_98fe47f5c7b5c509", "index": "infovqa_infovqa_000017_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "98fe47f5c7b5c50984efca4babeb2a224cadaab79ebcfae743a22b23c5aeb3a1", "original_question_id": "infovqa_000017", "question": "where should I eat if it is before 10:30 AM and I am not a police officer?\nGive a very brief answer.", "question_id": "infovqa_000017" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwhere should I eat if it is before 10:30 AM and I am not a police officer?\nGive a very brief answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_spatial_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "KFC." }
{ "all_answers": [ "KFC." ], "answer": "KFC.", "augmentation_type": "flip_v", "data_split": "train", "doc_id": "Infographic-VQA_98fe47f5c7b5c509", "index": "infovqa_infovqa_000018_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "98fe47f5c7b5c50984efca4babeb2a224cadaab79ebcfae743a22b23c5aeb3a1", "original_question_id": "infovqa_000018", "question": "where should I eat if I dont eat fish\nBe succinct.", "question_id": "infovqa_000018" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwhere should I eat if I dont eat fish\nBe succinct.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_spatial_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "IN-N-OUT Burger." }
{ "all_answers": [ "IN-N-OUT Burger." ], "answer": "IN-N-OUT Burger.", "augmentation_type": "rot180+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_98fe47f5c7b5c509", "index": "infovqa_infovqa_000019_rot180_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "98fe47f5c7b5c50984efca4babeb2a224cadaab79ebcfae743a22b23c5aeb3a1", "original_question_id": "infovqa_000019", "question": "Where should I eat if I live in westcoast and it is before 10:00PM\nShort answer required.", "question_id": "infovqa_000019" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhere should I eat if I live in westcoast and it is before 10:00PM\nShort answer required.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_spatial_rot180_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "18." }
{ "all_answers": [ "18." ], "answer": "18.", "augmentation_type": "rot270+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_98fe47f5c7b5c509", "index": "infovqa_infovqa_000020_rot270_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "98fe47f5c7b5c50984efca4babeb2a224cadaab79ebcfae743a22b23c5aeb3a1", "original_question_id": "infovqa_000020", "question": "How many eating joints have been mentioned here?\nAnswer briefly.", "question_id": "infovqa_000020" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many eating joints have been mentioned here?\nAnswer briefly.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot270_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "SONIC." }
{ "all_answers": [ "SONIC." ], "answer": "SONIC.", "augmentation_type": "rot180+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_98fe47f5c7b5c509", "index": "infovqa_infovqa_000021_rot180_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "98fe47f5c7b5c50984efca4babeb2a224cadaab79ebcfae743a22b23c5aeb3a1", "original_question_id": "infovqa_000021", "question": "Where can I go if I am in a car\nQuick response, please.", "question_id": "infovqa_000021" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhere can I go if I am in a car\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_spatial_rot180_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "6." }
{ "all_answers": [ "6." ], "answer": "6.", "augmentation_type": "rot180+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000022_rot180_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000022", "question": "LIver is a source of how many of the vitamins shown here?\nConcise answer only.", "question_id": "infovqa_000022" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nLIver is a source of how many of the vitamins shown here?\nConcise answer only.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot180_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Egg yolks." }
{ "all_answers": [ "Egg yolks." ], "answer": "Egg yolks.", "augmentation_type": "rot90+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000023_rot90_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000023", "question": "Which dairy item is a source of vitamin b9?\nAnswer briefly.", "question_id": "infovqa_000023" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich dairy item is a source of vitamin b9?\nAnswer briefly.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot90_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Liver." }
{ "all_answers": [ "Liver." ], "answer": "Liver.", "augmentation_type": "rot180", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000024_rot180", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000024", "question": "Which fish/meat item is a source of vitamin k?\nWrite a very short answer.", "question_id": "infovqa_000024" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich fish/meat item is a source of vitamin k?\nWrite a very short answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Shellfish." }
{ "all_answers": [ "Shellfish." ], "answer": "Shellfish.", "augmentation_type": "rot180", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000025_rot180", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000025", "question": "Which fish/meat item is a source of vitamin E?\nBe succinct.", "question_id": "infovqa_000025" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich fish/meat item is a source of vitamin E?\nBe succinct.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Avocado." }
{ "all_answers": [ "Avocado." ], "answer": "Avocado.", "augmentation_type": "rot180+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000026_rot180_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000026", "question": "Which fruit is a source of vitamin E?\nWrite a very short answer.", "question_id": "infovqa_000026" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich fruit is a source of vitamin E?\nWrite a very short answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Egg yolks." }
{ "all_answers": [ "Egg yolks." ], "answer": "Egg yolks.", "augmentation_type": "rot270+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000027_rot270_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000027", "question": "Apart from butter which dairy product is a source of vitamin e?\nShort answer required.", "question_id": "infovqa_000027" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nApart from butter which dairy product is a source of vitamin e?\nShort answer required.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot270_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Rice." }
{ "all_answers": [ "Rice." ], "answer": "Rice.", "augmentation_type": "rot180", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000028_rot180", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000028", "question": "Which grain is a source of vitamin b5 and b7?\nShort answer required.", "question_id": "infovqa_000028" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich grain is a source of vitamin b5 and b7?\nShort answer required.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Fortified milk." }
{ "all_answers": [ "Fortified milk." ], "answer": "Fortified milk.", "augmentation_type": "rot90+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000029_rot90_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000029", "question": "Apart from egg yolks, which dairy item is a source of vitamin D?\nGive a very brief answer.", "question_id": "infovqa_000029" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nApart from egg yolks, which dairy item is a source of vitamin D?\nGive a very brief answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot90_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Bread." }
{ "all_answers": [ "Bread." ], "answer": "Bread.", "augmentation_type": "rot180", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000030_rot180", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000030", "question": "Which grain item gets you vitamin b2?\nQuick response, please.", "question_id": "infovqa_000030" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich grain item gets you vitamin b2?\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Avocados." }
{ "all_answers": [ "Avocados." ], "answer": "Avocados.", "augmentation_type": "flip_v", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000031_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000031", "question": "Which fruit has vitamin b2?\nKeep it brief.", "question_id": "infovqa_000031" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich fruit has vitamin b2?\nKeep it brief.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "2." }
{ "all_answers": [ "2." ], "answer": "2.", "augmentation_type": "rot270+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000032_rot270_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000032", "question": "How many types of oils are listed here?\nGive a very brief answer.", "question_id": "infovqa_000032" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many types of oils are listed here?\nGive a very brief answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot270_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "5." }
{ "all_answers": [ "5." ], "answer": "5.", "augmentation_type": "rot180+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_007fba93483d51fa", "index": "infovqa_infovqa_000033_rot180_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "007fba93483d51faaec1d521b3f905ae4289825ab886c8f94d4235e723b54bca", "original_question_id": "infovqa_000033", "question": "How many meat/fish item are listed here?\nBe succinct.", "question_id": "infovqa_000033" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many meat/fish item are listed here?\nBe succinct.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot180_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "7 Years." }
{ "all_answers": [ "7 Years." ], "answer": "7 Years.", "augmentation_type": "rot90+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_c3b1e44ea4a03be9", "index": "infovqa_infovqa_000034_rot90_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "c3b1e44ea4a03be91280f4a5a858bfe109d2aaf7dcdb6498f0e2f92e62f2ce66", "original_question_id": "infovqa_000034", "question": "How long a swallowed gum will stuck in your tummy?\nConcise answer only.", "question_id": "infovqa_000034" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow long a swallowed gum will stuck in your tummy?\nConcise answer only.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_rot90_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "It strains your eyes, making them tired and achy." }
{ "all_answers": [ "It strains your eyes, making them tired and achy." ], "answer": "It strains your eyes, making them tired and achy.", "augmentation_type": "flip_v", "data_split": "train", "doc_id": "Infographic-VQA_c3b1e44ea4a03be9", "index": "infovqa_infovqa_000035_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "c3b1e44ea4a03be91280f4a5a858bfe109d2aaf7dcdb6498f0e2f92e62f2ce66", "original_question_id": "infovqa_000035", "question": "What is the truth of reading under low light?\nBe succinct.", "question_id": "infovqa_000035" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the truth of reading under low light?\nBe succinct.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Dance." }
{ "all_answers": [ "Dance." ], "answer": "Dance.", "augmentation_type": "flip_h", "data_split": "train", "doc_id": "Infographic-VQA_fa57652e54983974", "index": "infovqa_infovqa_000036_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "fa57652e5498397439a9985397ac402f71c2c910966d8f31d24e8f426d00240b", "original_question_id": "infovqa_000036", "question": "What is the third physical activity mentioned that helps to manage anxiety?\nWrite a very short answer.", "question_id": "infovqa_000036" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the third physical activity mentioned that helps to manage anxiety?\nWrite a very short answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Food." }
{ "all_answers": [ "Food." ], "answer": "Food.", "augmentation_type": "rot270", "data_split": "train", "doc_id": "Infographic-VQA_fa57652e54983974", "index": "infovqa_infovqa_000037_rot270", "infovqa_source": "Infographic-VQA", "original_image_filename": "fa57652e5498397439a9985397ac402f71c2c910966d8f31d24e8f426d00240b", "original_question_id": "infovqa_000037", "question": "What item can get delivered to save time?\nWrite a very short answer.", "question_id": "infovqa_000037" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat item can get delivered to save time?\nWrite a very short answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_temporal_rot270
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Clothes." }
{ "all_answers": [ "Clothes." ], "answer": "Clothes.", "augmentation_type": "flip_v", "data_split": "train", "doc_id": "Infographic-VQA_fa57652e54983974", "index": "infovqa_infovqa_000038_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "fa57652e5498397439a9985397ac402f71c2c910966d8f31d24e8f426d00240b", "original_question_id": "infovqa_000038", "question": "Which item should you discard if you stop using?\nAnswer briefly.", "question_id": "infovqa_000038" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich item should you discard if you stop using?\nAnswer briefly.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "4." }
{ "all_answers": [ "4." ], "answer": "4.", "augmentation_type": "rot180+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_fa57652e54983974", "index": "infovqa_infovqa_000039_rot180_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "fa57652e5498397439a9985397ac402f71c2c910966d8f31d24e8f426d00240b", "original_question_id": "infovqa_000039", "question": "How many tips are mentioned under the subheading FOOD?\nShort answer required.", "question_id": "infovqa_000039" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many tips are mentioned under the subheading FOOD?\nShort answer required.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot180_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "CHOCOLATE." }
{ "all_answers": [ "CHOCOLATE." ], "answer": "CHOCOLATE.", "augmentation_type": "rot270", "data_split": "train", "doc_id": "Infographic-VQA_83f2a6765986262c", "index": "infovqa_infovqa_000040_rot270", "infovqa_source": "Infographic-VQA", "original_image_filename": "83f2a6765986262c5f1c4e7bbe1d415b903247369f1591c2aedd526efde65517", "original_question_id": "infovqa_000040", "question": "Which type of sweet is tough to be paired with wine?\nBe succinct.", "question_id": "infovqa_000040" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich type of sweet is tough to be paired with wine?\nBe succinct.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot270
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "LIGHT RED." }
{ "all_answers": [ "LIGHT RED." ], "answer": "LIGHT RED.", "augmentation_type": "rot180+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_83f2a6765986262c", "index": "infovqa_infovqa_000041_rot180_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "83f2a6765986262c5f1c4e7bbe1d415b903247369f1591c2aedd526efde65517", "original_question_id": "infovqa_000041", "question": "Under what type of wine does Pinot Noir fall?\nWrite a very short answer.", "question_id": "infovqa_000041" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nUnder what type of wine does Pinot Noir fall?\nWrite a very short answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Green." }
{ "all_answers": [ "Green." ], "answer": "Green.", "augmentation_type": "flip_v", "data_split": "train", "doc_id": "Infographic-VQA_06102249da4c0e13", "index": "infovqa_infovqa_000042_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "06102249da4c0e13a30d5b78c4875440d13eac3fb416963ed77880407b3a4000", "original_question_id": "infovqa_000042", "question": "What is the colour of the sweater icon shown - green or red?\nGive a very brief answer.", "question_id": "infovqa_000042" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the colour of the sweater icon shown - green or red?\nGive a very brief answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "100 sq. feet." }
{ "all_answers": [ "100 sq. feet." ], "answer": "100 sq. feet.", "augmentation_type": "flip_h", "data_split": "train", "doc_id": "Infographic-VQA_bffe986afcb6992c", "index": "infovqa_infovqa_000043_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "bffe986afcb6992c15129a4d9a8e3ffce31a0da8818830b0ef9ec145f070dbef", "original_question_id": "infovqa_000043", "question": "What is the most common booth size?\nShort answer required.", "question_id": "infovqa_000043" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the most common booth size?\nShort answer required.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Boston." }
{ "all_answers": [ "Boston." ], "answer": "Boston.", "augmentation_type": "rot90", "data_split": "train", "doc_id": "Infographic-VQA_63949bc14b365841", "index": "infovqa_infovqa_000044_rot90", "infovqa_source": "Infographic-VQA", "original_image_filename": "63949bc14b3658411820617cda7c4dfb250bf5754c2b89f06d43a697228416ed", "original_question_id": "infovqa_000044", "question": "Where is the Northeastern university located as per the map?\nGive a very brief answer.", "question_id": "infovqa_000044" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhere is the Northeastern university located as per the map?\nGive a very brief answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_spatial_rot90
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "SAMUEL ADAMS." }
{ "all_answers": [ "SAMUEL ADAMS." ], "answer": "SAMUEL ADAMS.", "augmentation_type": "rot270", "data_split": "train", "doc_id": "Infographic-VQA_63949bc14b365841", "index": "infovqa_infovqa_000045_rot270", "infovqa_source": "Infographic-VQA", "original_image_filename": "63949bc14b3658411820617cda7c4dfb250bf5754c2b89f06d43a697228416ed", "original_question_id": "infovqa_000045", "question": "Where did Daniel Tewfik manage inventory?\nQuick response, please.", "question_id": "infovqa_000045" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhere did Daniel Tewfik manage inventory?\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_spatial_rot270
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "M." }
{ "all_answers": [ "M." ], "answer": "M.", "augmentation_type": "rot90+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_e0336c695b0f7693", "index": "infovqa_infovqa_000046_rot90_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "e0336c695b0f76930a39e250aac008ac51762c76e71aecbf5ca6b1ea23c4d244", "original_question_id": "infovqa_000046", "question": "Which is the letter written on the belt of Madonna?\nQuick response, please.", "question_id": "infovqa_000046" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich is the letter written on the belt of Madonna?\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot90_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "1991." }
{ "all_answers": [ "1991." ], "answer": "1991.", "augmentation_type": "rot180", "data_split": "train", "doc_id": "Infographic-VQA_e0336c695b0f7693", "index": "infovqa_infovqa_000047_rot180", "infovqa_source": "Infographic-VQA", "original_image_filename": "e0336c695b0f76930a39e250aac008ac51762c76e71aecbf5ca6b1ea23c4d244", "original_question_id": "infovqa_000047", "question": "In which year Madonnas Disco 1 was released?\nGive a very brief answer.", "question_id": "infovqa_000047" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nIn which year Madonnas Disco 1 was released?\nGive a very brief answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_temporal_rot180
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "97%." }
{ "all_answers": [ "97%." ], "answer": "97%.", "augmentation_type": "flip_h", "data_split": "train", "doc_id": "Infographic-VQA_f8eecf374ac56bf2", "index": "infovqa_infovqa_000048_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "f8eecf374ac56bf219c0495f6150be815ef023583b1846a99cfc71da6dc68ff7", "original_question_id": "infovqa_000048", "question": "What percent of the National GDP is not Direct GDP?\nBe succinct.", "question_id": "infovqa_000048" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat percent of the National GDP is not Direct GDP?\nBe succinct.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Mobile Optimisation." }
{ "all_answers": [ "Mobile Optimisation." ], "answer": "Mobile Optimisation.", "augmentation_type": "rot180", "data_split": "train", "doc_id": "Infographic-VQA_61cacc9e726228ac", "index": "infovqa_infovqa_000049_rot180", "infovqa_source": "Infographic-VQA", "original_image_filename": "61cacc9e726228ac07038605390a8c8bcd9f69e8d9488527c9cc91a22c526126", "original_question_id": "infovqa_000049", "question": "What is the 2nd benefit of mobile engagement\nKeep it brief.", "question_id": "infovqa_000049" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the 2nd benefit of mobile engagement\nKeep it brief.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Personalised Messages." }
{ "all_answers": [ "Personalised Messages." ], "answer": "Personalised Messages.", "augmentation_type": "rot90+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_61cacc9e726228ac", "index": "infovqa_infovqa_000050_rot90_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "61cacc9e726228ac07038605390a8c8bcd9f69e8d9488527c9cc91a22c526126", "original_question_id": "infovqa_000050", "question": "What is the 3rd benefit of mobile engagement\nGive a very brief answer.", "question_id": "infovqa_000050" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the 3rd benefit of mobile engagement\nGive a very brief answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot90_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Device integration." }
{ "all_answers": [ "Device integration." ], "answer": "Device integration.", "augmentation_type": "rot180+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_61cacc9e726228ac", "index": "infovqa_infovqa_000051_rot180_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "61cacc9e726228ac07038605390a8c8bcd9f69e8d9488527c9cc91a22c526126", "original_question_id": "infovqa_000051", "question": "What is the 4th benefit of mobile engagament\nConcise answer only.", "question_id": "infovqa_000051" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the 4th benefit of mobile engagament\nConcise answer only.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Personalised Messages." }
{ "all_answers": [ "Personalised Messages." ], "answer": "Personalised Messages.", "augmentation_type": "rot180+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_61cacc9e726228ac", "index": "infovqa_infovqa_000052_rot180_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "61cacc9e726228ac07038605390a8c8bcd9f69e8d9488527c9cc91a22c526126", "original_question_id": "infovqa_000052", "question": "What helps target different segments with ease\nBe succinct.", "question_id": "infovqa_000052" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat helps target different segments with ease\nBe succinct.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Smartphone." }
{ "all_answers": [ "Smartphone." ], "answer": "Smartphone.", "augmentation_type": "rot180+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_61cacc9e726228ac", "index": "infovqa_infovqa_000053_rot180_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "61cacc9e726228ac07038605390a8c8bcd9f69e8d9488527c9cc91a22c526126", "original_question_id": "infovqa_000053", "question": "What will replace desktop computers outside of the workplace\nConcise answer only.", "question_id": "infovqa_000053" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat will replace desktop computers outside of the workplace\nConcise answer only.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "4." }
{ "all_answers": [ "4." ], "answer": "4.", "augmentation_type": "rot90+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_2c89cfe4c0e8981c", "index": "infovqa_infovqa_000054_rot90_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "2c89cfe4c0e8981ce50c64e2c78dfe1bfacd8d3f888521e5d64a6cc8dca09461", "original_question_id": "infovqa_000054", "question": "How many people are not disabled in Australia considering five people?\nKeep it brief.", "question_id": "infovqa_000054" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many people are not disabled in Australia considering five people?\nKeep it brief.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot90_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "2 in 5 people." }
{ "all_answers": [ "2 in 5 people." ], "answer": "2 in 5 people.", "augmentation_type": "flip_h", "data_split": "train", "doc_id": "Infographic-VQA_2c89cfe4c0e8981c", "index": "infovqa_infovqa_000055_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "2c89cfe4c0e8981ce50c64e2c78dfe1bfacd8d3f888521e5d64a6cc8dca09461", "original_question_id": "infovqa_000055", "question": "What is the chance of disability for senior citizens?\nConcise answer only.", "question_id": "infovqa_000055" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the chance of disability for senior citizens?\nConcise answer only.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Female." }
{ "all_answers": [ "Female." ], "answer": "Female.", "augmentation_type": "rot90+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_2c89cfe4c0e8981c", "index": "infovqa_infovqa_000056_rot90_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "2c89cfe4c0e8981ce50c64e2c78dfe1bfacd8d3f888521e5d64a6cc8dca09461", "original_question_id": "infovqa_000056", "question": "Which category posses more percent of disability, male or female?\nKeep it brief.", "question_id": "infovqa_000056" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich category posses more percent of disability, male or female?\nKeep it brief.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot90_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "10%." }
{ "all_answers": [ "10%." ], "answer": "10%.", "augmentation_type": "rot90+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_2c89cfe4c0e8981c", "index": "infovqa_infovqa_000057_rot90_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "2c89cfe4c0e8981ce50c64e2c78dfe1bfacd8d3f888521e5d64a6cc8dca09461", "original_question_id": "infovqa_000057", "question": "What is the percentage of DYSLEXIA for Australians?\nShort answer required.", "question_id": "infovqa_000057" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the percentage of DYSLEXIA for Australians?\nShort answer required.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot90_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "45%." }
{ "all_answers": [ "45%." ], "answer": "45%.", "augmentation_type": "rot270", "data_split": "train", "doc_id": "Infographic-VQA_2c89cfe4c0e8981c", "index": "infovqa_infovqa_000058_rot270", "infovqa_source": "Infographic-VQA", "original_image_filename": "2c89cfe4c0e8981ce50c64e2c78dfe1bfacd8d3f888521e5d64a6cc8dca09461", "original_question_id": "infovqa_000058", "question": "What is the mental health condition percentage of Australians?\nWrite a very short answer.", "question_id": "infovqa_000058" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the mental health condition percentage of Australians?\nWrite a very short answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot270
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Mental health condition." }
{ "all_answers": [ "Mental health condition." ], "answer": "Mental health condition.", "augmentation_type": "rot270+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_2c89cfe4c0e8981c", "index": "infovqa_infovqa_000059_rot270_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "2c89cfe4c0e8981ce50c64e2c78dfe1bfacd8d3f888521e5d64a6cc8dca09461", "original_question_id": "infovqa_000059", "question": "For Australians percentage of Dyslexia or Mental health condition is more?\nAnswer briefly.", "question_id": "infovqa_000059" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nFor Australians percentage of Dyslexia or Mental health condition is more?\nAnswer briefly.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot270_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "5." }
{ "all_answers": [ "5." ], "answer": "5.", "augmentation_type": "rot270", "data_split": "train", "doc_id": "Infographic-VQA_2c89cfe4c0e8981c", "index": "infovqa_infovqa_000060_rot270", "infovqa_source": "Infographic-VQA", "original_image_filename": "2c89cfe4c0e8981ce50c64e2c78dfe1bfacd8d3f888521e5d64a6cc8dca09461", "original_question_id": "infovqa_000060", "question": "Out of six how many Australians are not affected with hearing loss?\nQuick response, please.", "question_id": "infovqa_000060" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nOut of six how many Australians are not affected with hearing loss?\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot270
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "4." }
{ "all_answers": [ "4." ], "answer": "4.", "augmentation_type": "rot180+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_3f37afcdc52e99e1", "index": "infovqa_infovqa_000061_rot180_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "3f37afcdc52e99e1f50f1421ddee685eb438b85211950c248862956957c22bdd", "original_question_id": "infovqa_000061", "question": "How many out of five where fined for speeding in Victoria ?\nWrite a very short answer.", "question_id": "infovqa_000061" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many out of five where fined for speeding in Victoria ?\nWrite a very short answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot180_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "More than 3500." }
{ "all_answers": [ "More than 3500." ], "answer": "More than 3500.", "augmentation_type": "rot90", "data_split": "train", "doc_id": "Infographic-VQA_3f37afcdc52e99e1", "index": "infovqa_infovqa_000062_rot90", "infovqa_source": "Infographic-VQA", "original_image_filename": "3f37afcdc52e99e1f50f1421ddee685eb438b85211950c248862956957c22bdd", "original_question_id": "infovqa_000062", "question": "How many fines are issued in Victoria each day?\nKeep it brief.", "question_id": "infovqa_000062" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many fines are issued in Victoria each day?\nKeep it brief.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot90
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Red." }
{ "all_answers": [ "Red." ], "answer": "Red.", "augmentation_type": "rot270+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_3f37afcdc52e99e1", "index": "infovqa_infovqa_000063_rot270_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "3f37afcdc52e99e1f50f1421ddee685eb438b85211950c248862956957c22bdd", "original_question_id": "infovqa_000063", "question": "What is the color of the car shown in the info graphic- yellow, orange, red or green?\nBe succinct.", "question_id": "infovqa_000063" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the color of the car shown in the info graphic- yellow, orange, red or green?\nBe succinct.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot270_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "1." }
{ "all_answers": [ "1." ], "answer": "1.", "augmentation_type": "rot180+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_3f37afcdc52e99e1", "index": "infovqa_infovqa_000064_rot180_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "3f37afcdc52e99e1f50f1421ddee685eb438b85211950c248862956957c22bdd", "original_question_id": "infovqa_000064", "question": "How many persons with Dollar symbol are depicted in the info graphic?\nBe succinct.", "question_id": "infovqa_000064" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many persons with Dollar symbol are depicted in the info graphic?\nBe succinct.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot180_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "4." }
{ "all_answers": [ "4." ], "answer": "4.", "augmentation_type": "flip_h", "data_split": "train", "doc_id": "Infographic-VQA_3f37afcdc52e99e1", "index": "infovqa_infovqa_000065_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "3f37afcdc52e99e1f50f1421ddee685eb438b85211950c248862956957c22bdd", "original_question_id": "infovqa_000065", "question": "How many person images are shown in grey color?\nQuick response, please.", "question_id": "infovqa_000065" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many person images are shown in grey color?\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "8." }
{ "all_answers": [ "8." ], "answer": "8.", "augmentation_type": "rot270", "data_split": "train", "doc_id": "Infographic-VQA_ee9e3df8641542fb", "index": "infovqa_infovqa_000066_rot270", "infovqa_source": "Infographic-VQA", "original_image_filename": "ee9e3df8641542fb1d988a9b011785cdd10a5bb95571717ebad7f39c3eb3e34a", "original_question_id": "infovqa_000066", "question": "How many election issues were reported by ABC?\nQuick response, please.", "question_id": "infovqa_000066" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many election issues were reported by ABC?\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot270
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Roads." }
{ "all_answers": [ "Roads." ], "answer": "Roads.", "augmentation_type": "flip_v", "data_split": "train", "doc_id": "Infographic-VQA_ee9e3df8641542fb", "index": "infovqa_infovqa_000067_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "ee9e3df8641542fb1d988a9b011785cdd10a5bb95571717ebad7f39c3eb3e34a", "original_question_id": "infovqa_000067", "question": "Which was the second last topic of the election issues?\nQuick response, please.", "question_id": "infovqa_000067" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhich was the second last topic of the election issues?\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "12." }
{ "all_answers": [ "12." ], "answer": "12.", "augmentation_type": "rot90+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_4a23f9471c0103dd", "index": "infovqa_infovqa_000068_rot90_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "4a23f9471c0103dd039d1eda9ec5c2c023c9c1c834704acc01cb09ce5a991e92", "original_question_id": "infovqa_000068", "question": "How many regions are shown in the map?\nAnswer briefly.", "question_id": "infovqa_000068" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many regions are shown in the map?\nAnswer briefly.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot90_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "5." }
{ "all_answers": [ "5." ], "answer": "5.", "augmentation_type": "flip_h", "data_split": "train", "doc_id": "Infographic-VQA_4a23f9471c0103dd", "index": "infovqa_infovqa_000069_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "4a23f9471c0103dd039d1eda9ec5c2c023c9c1c834704acc01cb09ce5a991e92", "original_question_id": "infovqa_000069", "question": "How many camels are shown in the info graphic?\nShort answer required.", "question_id": "infovqa_000069" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many camels are shown in the info graphic?\nShort answer required.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Internet." }
{ "all_answers": [ "Internet." ], "answer": "Internet.", "augmentation_type": "rot180+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_30bcb8a6e57cb28f", "index": "infovqa_infovqa_000070_rot180_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "30bcb8a6e57cb28f739049991741add0257275dbe20d12c618fe3956f8932e0a", "original_question_id": "infovqa_000070", "question": "What is the major reason for the wide usage of mobile phones?\nShort answer required.", "question_id": "infovqa_000070" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the major reason for the wide usage of mobile phones?\nShort answer required.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "6." }
{ "all_answers": [ "6." ], "answer": "6.", "augmentation_type": "rot90", "data_split": "train", "doc_id": "Infographic-VQA_30bcb8a6e57cb28f", "index": "infovqa_infovqa_000071_rot90", "infovqa_source": "Infographic-VQA", "original_image_filename": "30bcb8a6e57cb28f739049991741add0257275dbe20d12c618fe3956f8932e0a", "original_question_id": "infovqa_000071", "question": "How many different types of usage of Internet is depicted in the info graphic?\nWrite a very short answer.", "question_id": "infovqa_000071" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many different types of usage of Internet is depicted in the info graphic?\nWrite a very short answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot90
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "White." }
{ "all_answers": [ "White." ], "answer": "White.", "augmentation_type": "flip_h", "data_split": "train", "doc_id": "Infographic-VQA_30bcb8a6e57cb28f", "index": "infovqa_infovqa_000072_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "30bcb8a6e57cb28f739049991741add0257275dbe20d12c618fe3956f8932e0a", "original_question_id": "infovqa_000072", "question": "In which color Facebook logo is shown- black, white, red, blue?\nQuick response, please.", "question_id": "infovqa_000072" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nIn which color Facebook logo is shown- black, white, red, blue?\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Media." }
{ "all_answers": [ "Media." ], "answer": "Media.", "augmentation_type": "flip_v", "data_split": "train", "doc_id": "Infographic-VQA_30bcb8a6e57cb28f", "index": "infovqa_infovqa_000073_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "30bcb8a6e57cb28f739049991741add0257275dbe20d12c618fe3956f8932e0a", "original_question_id": "infovqa_000073", "question": "What is the text written inside the Facebook logo?\nAnswer briefly.", "question_id": "infovqa_000073" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat is the text written inside the Facebook logo?\nAnswer briefly.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "5." }
{ "all_answers": [ "5." ], "answer": "5.", "augmentation_type": "rot270+flip_h", "data_split": "train", "doc_id": "Infographic-VQA_925dbde74cfc199e", "index": "infovqa_infovqa_000074_rot270_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "925dbde74cfc199e45e636465fb98bc5c48997b66116b4f26c63e69dd01f4977", "original_question_id": "infovqa_000074", "question": "How many myths are listed in the info graphic?\nQuick response, please.", "question_id": "infovqa_000074" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many myths are listed in the info graphic?\nQuick response, please.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot270_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "10." }
{ "all_answers": [ "10." ], "answer": "10.", "augmentation_type": "flip_h", "data_split": "train", "doc_id": "Infographic-VQA_925dbde74cfc199e", "index": "infovqa_infovqa_000075_flip_h", "infovqa_source": "Infographic-VQA", "original_image_filename": "925dbde74cfc199e45e636465fb98bc5c48997b66116b4f26c63e69dd01f4977", "original_question_id": "infovqa_000075", "question": "What percent of Millennials does not own a smartphone?\nGive a very brief answer.", "question_id": "infovqa_000075" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat percent of Millennials does not own a smartphone?\nGive a very brief answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_flip_h
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "1." }
{ "all_answers": [ "1." ], "answer": "1.", "augmentation_type": "rot270+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_925dbde74cfc199e", "index": "infovqa_infovqa_000076_rot270_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "925dbde74cfc199e45e636465fb98bc5c48997b66116b4f26c63e69dd01f4977", "original_question_id": "infovqa_000076", "question": "How many millennial's chose not to live in metro capital cities?\nConcise answer only.", "question_id": "infovqa_000076" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nHow many millennial's chose not to live in metro capital cities?\nConcise answer only.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot270_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "60." }
{ "all_answers": [ "60." ], "answer": "60.", "augmentation_type": "flip_v", "data_split": "train", "doc_id": "Infographic-VQA_925dbde74cfc199e", "index": "infovqa_infovqa_000077_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "925dbde74cfc199e45e636465fb98bc5c48997b66116b4f26c63e69dd01f4977", "original_question_id": "infovqa_000077", "question": "What percentage of millennial's have not participated in the volunteer work?\nWrite a very short answer.", "question_id": "infovqa_000077" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat percentage of millennial's have not participated in the volunteer work?\nWrite a very short answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "35." }
{ "all_answers": [ "35." ], "answer": "35.", "augmentation_type": "rot180+flip_v", "data_split": "train", "doc_id": "Infographic-VQA_925dbde74cfc199e", "index": "infovqa_infovqa_000078_rot180_flip_v", "infovqa_source": "Infographic-VQA", "original_image_filename": "925dbde74cfc199e45e636465fb98bc5c48997b66116b4f26c63e69dd01f4977", "original_question_id": "infovqa_000078", "question": "What percentage of millennial's does not use smartphone to connect with others?\nKeep it brief.", "question_id": "infovqa_000078" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nWhat percentage of millennial's does not use smartphone to connect with others?\nKeep it brief.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_numerical_rot180_flip_v
visual_toolbox_v2
infovqa_augmented
{ "ground_truth": "Order, tip well, walk away." }
{ "all_answers": [ "Order, tip well, walk away." ], "answer": "Order, tip well, walk away.", "augmentation_type": "rot180", "data_split": "train", "doc_id": "Infographic-VQA_50b66d8c22dbf17b", "index": "infovqa_infovqa_000079_rot180", "infovqa_source": "Infographic-VQA", "original_image_filename": "50b66d8c22dbf17b2cd262ad9e0815efdb624e272a84e7aba28a1e6f9d7c2803", "original_question_id": "infovqa_000079", "question": "what do you do if you know what kind of beer you want\nGive a very brief answer.", "question_id": "infovqa_000079" }
[ { "content": "You are a helpful assistant.\n\n# Tools\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>\n{\"type\":\"function\",\"function\":{\"name\":\"image_zoom_in_tool\",\"description\":\"Zoom in on a specific region of an image by cropping it based on a bounding box (bbox) and an optional object label.\",\"parameters\":{\"type\":\"object\",\"properties\":{\"bbox_2d\":{\"type\":\"array\",\"items\":{\"type\":\"number\"},\"minItems\":4,\"maxItems\":4,\"description\":\"The bounding box of the region to zoom in, as [x1, y1, x2, y2], where (x1, y1) is the top-left corner and (x2, y2) is the bottom-right corner.\"},\"label\":{\"type\":\"string\",\"description\":\"The name or label of the object in the specified bounding box (optional).\"}},\"required\":[\"bbox\"]}}}\n</tools>\n\n# How to call a tool\nReturn a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call>\n\n**Example**: \n<tool_call> \n{\"name\": \"image_zoom_in_tool\", \"arguments\": {\"bbox_2d\": [10, 20, 100, 200], \"label\": \"the apple on the desk\"}} \n</tool_call>", "role": "system" }, { "content": "<image>\nwhat do you do if you know what kind of beer you want\nGive a very brief answer.\nThink first, call **image_zoom_in_tool** if needed, then answer. Format strictly as: <think>...</think> <tool_call>...</tool_call> (if tools needed) <answer>...</answer> ", "role": "user" } ]
infovqa_factual_rot180
visual_toolbox_v2
End of preview. Expand in Data Studio
README.md exists but content is empty.
Downloads last month
18