|
|
|
|
|
|
|
|
import json |
|
|
import pathlib |
|
|
|
|
|
import datasets |
|
|
import fsspec |
|
|
from datasets import DatasetInfo, Value, Features |
|
|
|
|
|
logger = datasets.logging.get_logger(__name__) |
|
|
|
|
|
_INFO = DatasetInfo( |
|
|
description='Automatically generated for wikitext (wikitext-103-raw-v1), split into 8 shards, detokenized.\n\nOriginal Description:\n The WikiText language modeling dataset is a collection of over 100 million tokens extracted from the set of verified\n Good and Featured articles on Wikipedia. The dataset is available under the Creative Commons Attribution-ShareAlike\n License.\n', |
|
|
citation='@misc{merity2016pointer,\n title={Pointer Sentinel Mixture Models},\n author={Stephen Merity and Caiming Xiong and James Bradbury and Richard Socher},\n year={2016},\n eprint={1609.07843},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n', |
|
|
homepage='https://blog.einstein.ai/the-wikitext-long-term-dependency-language-modeling-dataset/', |
|
|
license='Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)', |
|
|
version="1.0.0", |
|
|
features=Features.from_dict({'text': {'dtype': 'string', 'id': None, '_type': 'Value'}}), |
|
|
supervised_keys=None) |
|
|
|
|
|
|
|
|
class AutoDataset(datasets.GeneratorBasedBuilder): |
|
|
BUILDER_CONFIGS = [datasets.BuilderConfig()] |
|
|
|
|
|
def __init__(self, **kwargs): |
|
|
super().__init__(**kwargs) |
|
|
|
|
|
def _info(self): |
|
|
return _INFO |
|
|
|
|
|
@property |
|
|
def dataset_dir(self): |
|
|
return pathlib.Path(__file__).parent |
|
|
|
|
|
def _split_generators(self, dl_manager): |
|
|
metadata = json.load(open(dl_manager.download("metadata.json"), 'rt')) |
|
|
return [ |
|
|
datasets.SplitGenerator( |
|
|
name=split, |
|
|
gen_kwargs={"filepaths": dl_manager.download(split_metadata["files"])}, |
|
|
) |
|
|
for split, split_metadata in metadata["splits"].items() |
|
|
] |
|
|
|
|
|
def _generate_examples(self, filepaths): |
|
|
"""This function returns the examples in the raw (text) form by iterating on all the files.""" |
|
|
id_: int = 0 |
|
|
for filepath in filepaths: |
|
|
logger.info(f"Generating examples from {filepath}") |
|
|
with fsspec.open(filepath, mode="rt", compression="infer", encoding="utf-8") as f: |
|
|
for line in f: |
|
|
if line: |
|
|
example = json.loads(line) |
|
|
yield id_, example |
|
|
id_ += 1 |
|
|
|
|
|
|
|
|
if __name__ == "__main__": |
|
|
AutoDataset().download_and_prepare() |
