enwik8

ds = tfds.load('huggingface:enwik8/enwik8')

The dataset is based on the Hutter Prize (http://prize.hutter1.net) and contains the first 10^8 byte of Wikipedia

{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ds = tfds.load('huggingface:enwik8/enwik8-raw')

The dataset is based on the Hutter Prize (http://prize.hutter1.net) and contains the first 10^8 byte of Wikipedia

{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}