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

Fix spurious inputs failure when building Truncated variables #7328

Merged
merged 2 commits into from
May 25, 2024

Conversation

ricardoV94
Copy link
Member

@ricardoV94 ricardoV94 commented May 22, 2024

Description

Reintroduce dummy intermediate variables in implementation of TruncatedRV

Partially reverts 9d4a3d7 and 3888d53

The logprob derivation(s) in the icdf implementation of Truncated can duplicate nodes and cause spurious input variables to be marked as missing. We revert back to the old approach of replacing these by dummies so the graph above any inputs is hidden, and variables cannot be accidentally cloned/modified during logprob inference.

Related Issue

Checklist

Type of change

  • New feature / enhancement
  • Bug fix
  • Documentation
  • Maintenance
  • Other (please specify):

📚 Documentation preview 📚: https://pymc--7328.org.readthedocs.build/en/7328/

…edRV

Partially reverts 9d4a3d7 and 3888d53

The logprob derivation(s) in the icdf implementation of `Truncated` can duplicate nodes and cause spurious input variables to be marked as missing. We replace these by dummies so the graph above is hidden, and variables cannot be accidentally cloned/modified during logprob inference.
Copy link
Contributor

@tomicapretto tomicapretto left a comment

Choose a reason for hiding this comment

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

I don't have the capacity to judge the code modifications, but this fixes the problem that I reported on #7312.

@ricardoV94 ricardoV94 merged commit 19be124 into pymc-devs:main May 25, 2024
21 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
2 participants