Handle two-dimensional time embeddings - #1216
Open
fbonc wants to merge 2 commits into
Open
Conversation
Avoid a zero frequency denominator for the single-frequency case while preserving existing values for larger dimensions.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Issue
SinusoidalTimeEmbedding divided its frequency indices by half - 1. For dim=2, half is 1, so the only frequency was computed as 0/0 and every embedding contained NaN values.
Fix
Clamp to at least 1. This gives the single-frequency case a finite frequency of 1 while leaving the frequency schedule unchanged for all dimensions greater than 2.
Notes
Added regression tests tests/core/test_unified_embedding.py with finite-output coverage for dim=2 and a known-value check for dim=6.