This dataset contains Japanese captions for COCO images and their English translations.The format is CSV.
Dataset Structure
Data Fields
The data fields are the same among all lines.
filename(str): The name of the COCO image file
chatgpt text(str): The text generated by gpt-5-pro
gemini text(str): The text generated by gemini-3-pro-preview
grok text(str): The text generated by grok-4
claude text(str): The text generated by… See the full description on the dataset page: https://huggingface.co/datasets/baobab-trees/bao-val-coco-rating-cap.
