MultiCaptions-LARGE

# **MultiCaptions-LARGE** **MultiCaptions-LARGE** is a lightweight, multi-response image captioning dataset containing **20,592 image-caption pairs**, where each image is associated with **two complementary captions** generated by different vision-language models. - **response_1** contains a **dense, long-form caption** generated using the **Qwen3.5 Multimodal** model, providing detailed descriptions of scene composition, objects, attributes, spatial relationships, and visual context. - **response_2** contains a **short, precise caption** generated using the **LFM2.5 Vision-Language** model, offering concise descriptions suitable for lightweight captioning and retrieval tasks. The dataset is built primarily from **publicly available images**, which make up the majority of the input imagery, together with other publicly available datasets. > With a total size of only **632 MB**, **MultiCaptions-LARGE** is one of the most lightweight **20K+ multi-response image captioning datasets** available on the Hugging Face Hub. # Dataset Statistics | Property | Value | |----------|------:| | Number of Samples | **20,592** | | Total Size | **632 MB** | | Responses per Image | **2** | | License | Apache-2.0 | # Dataset Structure Each sample contains: | Column | Type | Description | |---------|------|-------------| | `image` | Image | Input image | | `response_1` | String | Dense caption generated by **Qwen3.5 Multimodal** | | `response_2` | String | Short caption generated by **LFM2.5 Vision-Language** | # Features - Dual-caption supervision for every image - Long-form dense captions - Short and precise captions - Lightweight dataset (632 MB) - More than 20K image-caption pairs - Suitable for instruction tuning and caption comparison - Apache-2.0 licensed # Citation If you use this dataset in your research or projects, please cite: ```bibtex @misc{prithiv_sakthi_2026, author = { Prithiv Sakthi }, title = { MultiCaptions-LARGE (Revision a764ac0) }, year = 2026, url = { https://huggingface.co/datasets/prithivMLmods/MultiCaptions-LARGE }, doi = { 10.57967/hf/9759 }, publisher = { Hugging Face } } ``` --- **Dataset by:** — https://huggingface.co/prithivMLmods