Adapter keeps track of the transform jacobians #419
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Thanks for the PR. Could you add tests for this behavior @Kucharssim? See e.g., |
|
Yes, on it! |
|
This was accidentally closed. We will investigate how to restore the branch and reopen PRs. |
LarsKue
requested changes
Apr 23, 2025
LarsKue
approved these changes
Apr 25, 2025
There was a problem hiding this comment.
Pull Request Overview
This PR enhances the adapter functionality by adding support for tracking the log determinant of the Jacobians across transforms. Key changes include:
- Introducing log_det_jac methods for several transforms (e.g. standardize, sqrt, scale, constrain, etc.).
- Modifying the adapter’s forward and inverse methods to optionally return the log determinant of the Jacobian.
- Adding comprehensive tests to verify the correct computation and exception handling for the Jacobian tracking.
Reviewed Changes
Copilot reviewed 18 out of 18 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
| tests/test_adapters/test_adapters.py | Updated tests to cover additional keys and jacobian tracking tests. |
| tests/test_adapters/conftest.py | Introduced new fixtures that test the log_det_jac functionality. |
| bayesflow/approximators/continuous_approximator.py | Modified forward method to incorporate and apply the change‐of‐variables formula. |
| bayesflow/adapters/transforms/* | Added or updated log_det_jac methods across multiple transform files. |
| bayesflow/adapters/adapter.py | Updated forward and inverse methods to support the log_det_jac flag. |
|
Ok, I think I addressed everything now. If you are happy with this, feel free to merge |
# for free
to join this conversation on GitHub.
Already have an account?
# to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
fixes #245
I have not implemented it for
NumpyTransformbecause I was not sure about implementation w.r.t. serialization. I tried to make sure that if we cannot keep track of the jacobians, we raise an error rather than returning an incorrect output.