EleutherAI / the-pile

MIT License
1.44k stars 122 forks source link

failed to download stackexchange #97

Open sangmichaelxie opened 2 years ago

sangmichaelxie commented 2 years ago

Running the python the_pile/pile.py --force_download command, I get the following error when downloading stackexchange:

Exception: Failed to download components/stackexchange/stackexchange_dataset.tar from any source

It seems like stackenchange_dataset.tar does download and almost fully finish, but when the code tries to unzip the download, the tar isn't there.

zinuzian commented 1 year ago

faced same issue here :(