Datasets:

Silviase
/

JA-VG-VQA-500-EN

image_id int64 11 2.42M	url stringlengths 52 59	width int64 227 1.28k	height int64 223 1.28k	coco_id float64 4.19k 580k ⌀	flickr_id float64 4.29M 10.3B ⌀	qas listlengths 8 8	image dict	language stringclasses 1 value	standard_prompt stringclasses 1 value
2,338,292	https://cs.stanford.edu/…100K/2338292.jpg	375	500	null	5,279,266,151	[ { "a_objects": [], "answer": "車のある屋外の場所", "q_objects": [], "qa_id": 10659136, "question": "場所はどこになりますか？" }, { "a_objects": [], "answer": "昼間", "q_objects": [], "qa_id": 10659137, "question": "時間帯はいつごろですか？" }, { "a_objects": [], "answer": "晴れている", "q_object...	{ "bytes": [ 255, 216, 255, 224, 0, 16, 74, 70, 73, 70, 0, 1, 1, 1, 0, 72, 0, 72, 0, 0, 255, 219, 0, 67, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3...	en	Answer briefly to the questions about the image.
2,351,612	https://cs.stanford.edu/…100K/2351612.jpg	500	374	451,976	6,949,585,090	[{"a_objects":[],"answer":"池のほとりです。","q_objects":[],"qa_id":10719456,"question":"こ(...TRUNCATED)	{"bytes":"/9j/4AAQSkZJRgABAQEASABIAAD//gAMQXBwbGVNYXJrCv/iDFhJQ0NfUFJPRklMRQABAQAADEhMaW5vAhAAAG1udH(...TRUNCATED)	en	Answer briefly to the questions about the image.
2,351,367	https://cs.stanford.edu/…100K/2351367.jpg	500	375	73,730	218,494,045	[{"a_objects":[],"answer":"１台","q_objects":[],"qa_id":10163616,"question":"ベッドは何台あ(...TRUNCATED)	{"bytes":"/9j/4AAQSkZJRgABAQEAZABkAAD/2wBDAAEBAQEBAQEBAQEBAQECAgMCAgICAgQDAwIDBQQFBQUEBAQFBgcGBQUHBg(...TRUNCATED)	en	Answer briefly to the questions about the image.
2,357,482	https://cs.stanford.edu/…100K/2357482.jpg	397	500	null	5,350,454,182	[{"a_objects":[],"answer":"アメリカ","q_objects":[],"qa_id":10135120,"question":"画像に写っ(...TRUNCATED)	{"bytes":"/9j/4AAQSkZJRgABAQEASABIAAD/4gxYSUNDX1BST0ZJTEUAAQEAAAxITGlubwIQAABtbnRyUkdCIFhZWiAHzgACAA(...TRUNCATED)	en	Answer briefly to the questions about the image.
2,374,342	https://cs.stanford.edu/…100K/2374342.jpg	500	332	322,149	5,634,948,997	[{"a_objects":[],"answer":"２頭です。","q_objects":[],"qa_id":10071688,"question":"動物の姿(...TRUNCATED)	{"bytes":"/9j/4AAQSkZJRgABAQEASABIAAD/4gUoSUNDX1BST0ZJTEUAAQEAAAUYYXBwbAIgAABzY25yUkdCIFhZWiAH0wAHAA(...TRUNCATED)	en	Answer briefly to the questions about the image.
2,384,661	https://cs.stanford.edu/…0K_2/2384661.jpg	375	500	337,062	6,987,950,916	[{"a_objects":[],"answer":"3つ","q_objects":[],"qa_id":10439520,"question":"標識はいくつあ(...TRUNCATED)	{"bytes":"/9j/4AAQSkZJRgABAQEASABIAAD/2wBDAAEBAQEBAQEBAQEBAQECAgMCAgICAgQDAwIDBQQFBQUEBAQFBgcGBQUHBg(...TRUNCATED)	en	Answer briefly to the questions about the image.
2,327,882	https://cs.stanford.edu/…100K/2327882.jpg	500	464	367,357	6,991,901,328	[{"a_objects":[],"answer":"3人","q_objects":[],"qa_id":10011464,"question":"全部で何人の人(...TRUNCATED)	{"bytes":"/9j/4AAQSkZJRgABAQEAYABgAAD/2wBDAAEBAQEBAQEBAQEBAQECAgMCAgICAgQDAwIDBQQFBQUEBAQFBgcGBQUHBg(...TRUNCATED)	en	Answer briefly to the questions about the image.
2,410,754	https://cs.stanford.edu/…0K_2/2410754.jpg	500	375	542,231	378,941,452	[{"a_objects":[],"answer":"女性","q_objects":[],"qa_id":10608328,"question":"この人の性別は(...TRUNCATED)	{"bytes":"/9j/4AAQSkZJRgABAQEASABIAAD/4gUoSUNDX1BST0ZJTEUAAQEAAAUYYXBwbAIgAABzY25yUkdCIFhZWiAH0wAHAA(...TRUNCATED)	en	Answer briefly to the questions about the image.
2,319,692	https://cs.stanford.edu/…100K/2319692.jpg	500	375	194,885	7,992,782,273	[{"a_objects":[],"answer":"３つ","q_objects":[],"qa_id":10384776,"question":"ベッドはいくつ(...TRUNCATED)	{"bytes":"/9j/4AAQSkZJRgABAQEASABIAAD/2wBDAAEBAQEBAQEBAQEBAQECAgMCAgICAgQDAwIDBQQFBQUEBAQFBgcGBQUHBg(...TRUNCATED)	en	Answer briefly to the questions about the image.
2,346,800	https://cs.stanford.edu/…100K/2346800.jpg	500	333	219,313	4,042,477,662	[{"a_objects":[],"answer":"黒色","q_objects":[],"qa_id":10555936,"question":"背景はどんな色(...TRUNCATED)	{"bytes":"/9j/4AAQSkZJRgABAQEA8ADwAAD/4gxYSUNDX1BST0ZJTEUAAQEAAAxITGlubwIQAABtbnRyUkdCIFhZWiAHzgACAA(...TRUNCATED)	en	Answer briefly to the questions about the image.

End of preview. Expand in Data Studio

Ja Vg Vqa 500 (English Translation)

This is an English translation of the SakanaAI/JA-VG-VQA-500 dataset, translated using plamo-translate.

Dataset Description

This dataset is part of the llm-jp-eval-mm benchmark suite. The original Japanese questions and answers have been translated to English while preserving the visual content.

Translation Details

Translation Model: pfnet/plamo-translate
Fields Translated: question, answer
Original Language: Japanese
Target Language: English

Usage

You can use this dataset with the llm-jp-eval-mm framework by specifying the language parameter:

from eval_mm.tasks import JaVgVqa500

# Use English version
task = JaVgVqa500(language="en")

# Use original Japanese version
task = JaVgVqa500(language="ja")

Downloads last month: 4

Data Sourcing report

powered

by Spawning.ai

No elements in this dataset have been identified as either opted-out, or opted-in, by their creator.