References:
plain_text
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:lm1b/plain_text')
- Description:
A benchmark corpus to be used for measuring progress in statistical language modeling. This has almost one billion words in the training data.
- License: No known license
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'test' |
306688 |
'train' |
30301028 |
- Features:
{
"text": {
"dtype": "string",
"id": null,
"_type": "Value"
}
}