Fix #506 (missing logits_projector serialization) #507
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Outdated suggestions cannot be applied.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.
Minimal fix providing missing serialization of the final logits_projector layer handled by the
ModelComparisonApproximatorclass itself. Both the weights and the predictions of an original and a reloaded approximator of the example in #506 are now exactly equal. Please let me know if any further adaptations for other cases are needed.As a more general discussion point: As there were some issues related to saving/loading recently, should we add some tests (e.g., using tempfiles) for the approximators that ensure that the weights and predictions of an original and a reloaded approximator match?