A curated image captioning dataset containing 1,000 image and caption pairs. Each sample consists of an image paired with a detailed natural language description, making it suitable for training and evaluating image captioning and vision-language models.

Features

1,000 image-caption pairs Detailed natural language captions High-quality descriptive annotations Suitable for image captioning and multimodal learning

Data Format

{… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/DeepCaption-1K-Qwen35.