Skip to content

Fix host URL in The Pile datasets#3627

Merged
albertvillanova merged 5 commits into
masterfrom
fix-3626
Feb 14, 2022
Merged

Fix host URL in The Pile datasets#3627
albertvillanova merged 5 commits into
masterfrom
fix-3626

Conversation

@albertvillanova

Copy link
Copy Markdown
Member

This PR fixes the host URL in The Pile datasets, once they have mirrored their data in another server.

Fix #3626.

@mariosasko

Copy link
Copy Markdown
Collaborator

We should also update the bookcorpusopen download url (see #3561) , no?

@albertvillanova

albertvillanova commented Jan 26, 2022

Copy link
Copy Markdown
Member Author

For the_pile_openwebtext2 and the_pile_stack_exchange I did not regenerate the JSON files, but instead I just changed the download_checksums URL.

@TrentBrick

Copy link
Copy Markdown

Seems like the mystic URL is now broken and the original should be used.

@TrentBrick

Copy link
Copy Markdown

Also if I git clone and edit the repo or reset it before this PR it is still trying to pull using mystic? Why is this?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

The Pile cannot connect to host

3 participants