fix(higgs): strip delay-pattern BOS/EOS columns before codec decode#843
Closed
Thorfy wants to merge 1 commit into
Closed
fix(higgs): strip delay-pattern BOS/EOS columns before codec decode#843Thorfy wants to merge 1 commit into
Thorfy wants to merge 1 commit into
Conversation
After revert_delay_pattern, the EOS ramp leaves mixed marker+code columns. The old [:, 1:-1] trim only removes all-marker edges; surviving stream ids were clipped to codebook_size-1 (1023) and decoded as a systematic breath at utterance end. Drop any column that still carries stream BOS/EOS instead. Co-authored-by: Cursor <cursoragent@cursor.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
revert_delay_pattern.[:, 1:-1]only drops all-marker edges; survivingaudio_stream_eos_idvalues wereclip'd to 1023 and decoded as a systematic end breath.strip_stream_marker_columns(oneshot + overlap-add decode paths).Fixes #842
Why this is not the Qwen3 fix
Qwen3 trailing breath is late
codec_eossampling after PR 695 (tracked separately). Higgs is a post-revert token hygiene bug in the delay-pattern pipeline.Test plan
python -m unittest mlx_audio.tts.tests.test_higgs_audio.TestStripStreamMarkerColumnsstream=Trueoverlap-add path: same check on final chunk