-
Notifications
You must be signed in to change notification settings - Fork 10
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
Pass hyper params from client #21
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looking great in overall!
Just some comments/questions:
- regarding the model args, why not pop the
model
args in the client instead and keep the server untouched? Then we'll have one less change to make when we have the feature enabled. - can we add some test cases, concerning the presence of
tabpfn_config
? Just a very minimal safe guard to ensure that we taketabpfn_config
Tests: yeah, good point. I will add a test ensuring the defaults are send. "model" arg: at some point we will need this model arg anyways, as we will train new models and want users to be able to choose, so I thought why not now. I should make it backward compatible on the backend, though. |
i added a test and updated it to follow the new API. Can I merge @liam-sbhoo ? |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Some issues on the validation_link
.
Change Description
Brief (a few bullet points describing your changes, use full sentences and try to link lines in the code whenever needed)
In this PR, I start passing all arguments to the server. To make this simpler I removed the local tabpfn functionality for now but kept the main code s.t. we can easily re-enable it later on. This was simpler, because the local tabpfn has very different parameters from the server one.
DEPENDENCY: this PR depends on https://github.com/automl/tabpfn-server/pull/48/ to filter out the "model" kwarg
Standard Qs (leave questions that do not apply blank)
If you broke behavior: Please describe what behavior you broke and how you inform people to not get stuck trying to use the old behavior.
yes, i removed local
If you used new dependencies: Did you add them to
requirements.txt
?no
Who did you ping on Mattermost to review your PR? Please ping that person again whenever you are ready for another review.
@liam-sbhoo @anshulg954 @davidotte
Please do not mark comments/conversations as resolved unless you are the assigned reviewer. This helps maintain clarity during the review process.