Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

force download when previous file is obsolete #1939

Merged
merged 2 commits into from
Oct 9, 2023

Conversation

AndreaFrancis
Copy link
Contributor

@AndreaFrancis AndreaFrancis commented Oct 5, 2023

Related to #1443 for cause_exception="OSError"
For https://datasets-server.huggingface.co/search?dataset=xnli&config=all_languages&split=train&offset=0&length=1&query=language


OSError: Consistency check failed: file should be of size 301736502 but has size 293192246 (0002.parquet).
We are sorry for the inconvenience. Please retry download and pass `force_download=True, resume_download=False` as argument.
If the issue persists, please let us know by opening an issue on https://github.com/huggingface/huggingface_hub.

See that there is an error in hf_hub_download, It looks like the library tried to compare the file with a previous one in the local cache. Maybe the data changed or there was a previously downloaded obsolete/corrupted file.
I think we should force-download each time to always have a refreshed parquet file.

@AndreaFrancis AndreaFrancis marked this pull request as ready for review October 6, 2023 18:08
@AndreaFrancis
Copy link
Contributor Author

I will merge this PR since looks like the e2e failed check is not related to this code.
I was able to run the tests locally
Screenshot from 2023-10-09 08-32-05

@AndreaFrancis AndreaFrancis merged commit 54ebe02 into main Oct 9, 2023
3 of 4 checks passed
@AndreaFrancis AndreaFrancis deleted the force-download-hf-hub-download branch October 9, 2023 12:33
@severo
Copy link
Collaborator

severo commented Oct 9, 2023

I will merge this PR since looks like the e2e failed check is not related to this code.

https://twitter.com/tmcw/status/1651646574873681925 <= we're not the only ones with half-broken e2e tests... But we have to fix 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.

2 participants