COCO-Mini-DeepCaption-10K: 10,000 Images with Detailed Synthetic Captions

COCO-Mini-DeepCaption-10K is a dense image captioning dataset built from a 10,000-image subset of the COCO dataset. It pairs these images with long-form synthetic captions generated using the Qwen3.5 multimodal model. The dataset was created by prithivMLmods and was last updated on July 6, 2026.

Use Cases

• Train dense image captioning models based on the detailed, long-form synthetic captions.
• Benchmark vision-language models on the quality of scene composition and subject attribute descriptions.
• Fine-tune generative models for high-fidelity image-to-text tasks using the Qwen3.5-generated captions.
• Research synthetic data quality for computer vision applications based on the described captioning pipeline.

Strengths

• Contains 10,000 images, providing a substantial base for model training.
• Captions are described as long-form, detailed, and high-fidelity, focusing on scene composition and subject attributes.
• Uses a dedicated Qwen3.5 captioning pipeline, suggesting a structured generation process.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect bias inherent to the source COCO dataset.

Provenance

Source

prithivMLmods on Hugging Face.

Collection Method

Built from a subset of the COCO dataset, with captions generated by the Qwen3.5 multimodal model.

Freshness

Last updated 2026-07-06 03:58:04; freshness should be verified.