Datasets:

nyanko7

/

danbooru2023

like 141

We made a supplement dataset for danbooru, sync all the newest images daily: https://huggingface.co/datasets/deepghs/danbooru_newest

And we are developing some new techs to quickly download tons of images with given ids: https://github.com/deepghs/cheesechaser

It's a wip project but make sense. Here is an example:

from cheesechaser.datapool import DanbooruNewestDataPool

pool = DanbooruNewestDataPool()

# download danbooru images, to directory /data/exp2
pool.batch_download_to_directory(
    resource_ids=[
        # old images, download from nyanko7/danbooru2023
        *range(2010000, 2010100),

        # newest images, download from deepghs/danbooru_newest
        *range(7508529, 7508629)
    ],
    dst_dir='/data/exp2',
    max_workers=12,
)