launch-calcium/e621-2024

--- license: mit task_categories: - image-classification - image-to-image - text-to-image pretty_name: e621-2024 size_categories: - 1M<n<10M viewer: true --- # e621-2024 e621-2024 is a large-scale furry image dataset retrieved from e621, a mature furry imageboard. This dataset is heavily inspired by the [danbooru2023](https://huggingface.co/datasets/nyanko7/danbooru2023) dataset. Similar to danbooru2023, the images in this dataset are bucketed into 1000 subdirectories (0000-0999), which is the E621 ID modulo 1000 (so all images in 0999/ have an ID ending in '999'). Currently there is no loading script for this dataset, so loading it with HuggingFace datasets is not supported. This dataset was prepared using the [daily database export](https://e621.net/db_export/) for 2024-04-07. A copy of that export in both the original .csv.gz format and parquet is in the metadata folder.