openwebtext2
openwebtext2 copied to clipboard
EleutherAI
Hello, I have been trying to download the datasets, but both links are not working. Could someone please take a look at the downloadable links and implement a fix for...
The pushshift.pushshift_to_sqlite method passes the arguments to best_download.download_file in a wrong order, and the code crashes. Hence, the dataset is not reproducible without this modification.
The link of openwebtext2 seems failed to open for download, can someone help to check it? page: https://openwebtext2.readthedocs.io/en/latest/index.html#download-plug-and-play-version link: https://mystic.the-eye.eu/public/AI/pile_preliminary_components/openwebtext2.jsonl.zst.tar (failed to open)
