Dataset · trojblue
Danbooru anime art metadata and tags (2025 snapshot)
trojblue/danbooru2025-metadata
Fresh dump of tags and metadata for millions of images from the Danbooru imageboard in parquet format.
What's inside
Structured tabular and text metadata of user anime illustrations collected from the Danbooru imageboard up to April 2025 (latest post ID is 9,158,800). The volume is 1 to 10 million records in English and Japanese, all packaged in parquet format. Compared to older versions, tag history tracking and coverage of older posts have been improved.
What you can do with it
- Training auto-taggers and anime image classifiers
- Preparing text prompts for training LoRA and art generation models
- Analysis of tag evolution and popularity in the anime community
- Tasks
- Text → image, Image classification
- Languages
- English, Japanese
- Size
- 1M–10M rows
- Downloads this month
- 1.3K
- Likes
- 39
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("trojblue/danbooru2025-metadata")