Гутенберг_время

ds = tfds.load('huggingface:gutenberg_time/gutenberg')

A clean data resource containing all explicit time references in a dataset of 52,183 novels whose full text is available via Project Gutenberg.

{
    "guten_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "hour_reference": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "time_phrase": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "is_ambiguous": {
        "dtype": "bool_",
        "id": null,
        "_type": "Value"
    },
    "time_pos_start": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "time_pos_end": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "tok_context": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}