dataset_infos.json

{
    "default": {
        "description": "Digital Peter: recognition of Peter the Great's manuscripts",
        "features": {
            "image": {"decode": true, "id": null, "_type": "Image"}
        },
        "splits": {"test": {
            "name": "train",
            "num_bytes": 100259,
            "num_examples": 662,
            "dataset_name": "images"
        }}
    }
}