Aesthetic Image Captions 10K for Diffusion Model Training

A curated set of 10,000 high-resolution image-caption pairs for training and research. Images were sourced from Pexels, and captions were generated with JoyCaption before being cleaned for use. The dataset was created by edwixx and was last updated on 2026-05-16.

Use Cases

• Train diffusion model LoRAs based on curated high-resolution image-caption pairs.
• Conduct research on image generation based on aesthetic image data.
• Perform experiments on caption quality based on the cleaned, concise captions.

Strengths

• Contains 10,000 high-resolution image-caption pairs.
• Captions were cleaned into concise sidecars specifically for diffusion training.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Freshness should be verified; last updated 2026-05-16.

Provenance

Source

Pexels (images), JoyCaption (initial captions).

Collection Method

Images sourced from Pexels, captions generated with JoyCaption and then cleaned.

Freshness

2026-05-16