Fix ACE-Step 1.5: max_tokens typo and lyrics embedding truncation#12529
Open
fishmongr wants to merge 1 commit intoComfy-Org:masterfrom
Open
Fix ACE-Step 1.5: max_tokens typo and lyrics embedding truncation#12529fishmongr wants to merge 1 commit intoComfy-Org:masterfrom
fishmongr wants to merge 1 commit intoComfy-Org:masterfrom
Conversation
Two bugs in ace15.py encode_token_weights(): 1. max_tokens parameter received min_tokens value: `max_tokens=lm_metadata["min_tokens"]` → `max_tokens=lm_metadata["max_tokens"]` This caused the LM to always generate minimum-length audio codes regardless of the requested duration. 2. lyrics_embeds[:, 0] discarded the full lyric sequence, passing only the first token embedding to the diffusion model's lyric encoder. Changed to pass the full lyrics_embeds tensor for proper lyrics conditioning. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
fishmongr
requested review from
Kosinkadink,
comfyanonymous and
guill
as code owners
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Two bugs in
comfy/text_encoders/ace15.pyencode_token_weights():max_tokensreceivesmin_tokensvalue:max_tokens=lm_metadata["min_tokens"]should bemax_tokens=lm_metadata["max_tokens"]. This causes the LM to always generate minimum-length audio codes regardless of the requested duration.lyrics_embeds[:, 0]truncates lyric sequence: Only the first token embedding was passed to the diffusion model's lyric encoder instead of the full sequence. This effectively discards all lyrics conditioning beyond the first token.Test plan
🤖 Generated with Claude Code