Fix uninformative val_loss - #60
Merged
themattinthehatt merged 2 commits intoAug 6, 2026
Merged
Conversation
…nd at validation At validation, _sample_background returns None. The renderer turns None into a white background while _prepare_target composites the target on black, so the unmasked perceptual term sees a whole-frame background mismatch and val_loss is inflated to a near-constant value. Return a fixed black background at eval so the render and the composited target share it. Training (random_background) is unchanged.
themattinthehatt
approved these changes
Aug 6, 2026
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 #59
At validation, _sample_background returns None. The renderer turns None into a white background while _prepare_target composites the target on black, so the unmasked perceptual term sees a whole-frame background mismatch and val_loss is inflated to a near-constant value.
Fixed by returning a fixed black background at eval so the render and the composited target share it. Training (random_background) is unchanged.