This dataset is a synthetic VQA dataset for Manga. Using images from Manga109, we generated VQA samples by providing both images and its text annotations to GPT-4o (gpt-4o-2024-11-20).
The dataset contains a total of 41,895 synthetic VQA samples from 8,379 images.
License
Creative Commons Attribution 4.0 License; and it should abide by the OpenAI terms of use
