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

Support copying posterior caches (and thereby float32 predictions from float64 models) #891

Draft
wants to merge 1 commit into
base: develop
Choose a base branch
from

Conversation

uri-granta
Copy link
Collaborator

Related issue(s)/PRs:

Summary

...

Fully backwards compatible: yes / no

PR checklist

  • The quality checks are all passing
  • The bug case / new feature is covered by tests
  • Any new features are well-documented (in docstrings or notebooks)

@uri-granta uri-granta changed the title Support copying posterior caches Support copying posterior caches (and thereby float32 predictions from float64 models) Jan 20, 2025
Comment on lines +370 to +372
npt.assert_allclose(f_mean, f_mean32, rtol=1e-6)
assert f_var32.dtype is tf.float32
npt.assert_allclose(f_var, f_var32, rtol=1e-6)
Copy link
Collaborator

Choose a reason for hiding this comment

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

my assumption is that accuracy will depend a lot on the problem, here we are testing on a simple function and its not surprising we get pretty much the same value - have you tested it on more complex problems of higher dimensionality?

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

Good call. Can you suggest a problem I can easily test it on?

Copy link
Collaborator

Choose a reason for hiding this comment

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

lets start with more complex functions that we have in our synthetic problems - we have some where we can specify arbitrary dimensionality as well I think

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