Dataset · KBlueLeaf
Danbooru2023 Metadata in SQLite
KBlueLeaf/danbooru2023-metadata-database
An SQLite database of Danbooru2023 tags and posts with a latest ID of 7,866,491 for generative models.
What's inside
This is an SQLite database containing Danbooru2023 post and tag metadata; the latest ID is 7,866,491, which is newer than the nyanko7 version. A Peewee ORM config is included, showing post-tag relationships. The source data is the official post information dump; this is a metadata database, not an image collection.
What you can do with it
- Training text-to-image models on Danbooru tags
- Building a filtered image dataset
- Analyzing tags and posts
- Tasks
- Image classification, Text → image, Image → text, Image → image, Text retrieval, Text generation
- Languages
- English, Japanese
- Size
- 1M–10M rows
- Downloads this month
- 522
- Likes
- 84
- Access
- on request on Hugging Face
Models trained on it
- KBlueLeaf/TIPO-v2.1-1B-A200MText generation 21.7K
- KBlueLeaf/TIPO-500M-ftText generation 18.6K
- KBlueLeaf/TIPO-200M-ft2Text generation 8.3K
- KBlueLeaf/TIPO-500MText generation 7.2K
- KBlueLeaf/TIPO-200MText generation 5K
- KBlueLeaf/TIPO-100MText generation 2K
- KBlueLeaf/TIPOv2-1B-A200MText generation 1.4K
- KBlueLeaf/TIPO-200M-ftText generation 656
How to load it
from datasets import load_dataset
ds = load_dataset("KBlueLeaf/danbooru2023-metadata-database")