Skip to content

Conversation

@ncclementi
Copy link
Contributor

This PR documents how to run a simple case of cudf.pandas accelerator on Modal, using their @app.function decorator.

I had to document a few caveats regarding the python installation, and the limitation on using the default RMM mode.

Out of the scope of this PR:

  • Document Modal NB, it'll be done in a followup.

cc: @charlesfrye for visibility if you want to take a look at the docs.

@ncclementi ncclementi requested a review from a team as a code owner January 8, 2026 21:03
@charlesfrye
Copy link

lgtm! thanks naty

@charlesfrye
Copy link

Maybe mention that Modal includes $30/month in free compute credits for all accounts, which is enough for ~10 hours of H100 time at current rates?

Copy link
Member

@jacobtomlinson jacobtomlinson left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @ncclementi, I left a few comments.

@charlesfrye I don't think we can mention the amount of included credits or specific GPU models because those things may change over time. I would be happy for us to write something more evergreen like mentioning modal has a free-tier and link to your pricing page, provided there is high confidence there will always be a free-tier.

@charlesfrye
Copy link

@jacobtomlinson yep, there will always be a free tier with included credits

@jacobtomlinson
Copy link
Member

/merge

@jacobtomlinson jacobtomlinson merged commit f3487c2 into rapidsai:main Jan 13, 2026
3 checks passed
@ncclementi ncclementi deleted the modal-docs branch January 13, 2026 17:20
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.

Document RAPIDS on Modal

3 participants