Skip to content

Fix audio decoding: preserve long dtype for token indices #19

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
lmangani merged 2 commits into from
Jan 30, 2026
Merged

Fix audio decoding: preserve long dtype for token indices #19

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
684b997
Initial plan
Copilot Jan 30, 2026
c187b41
Fix tensor dtype issue in audio decoding - keep frames as long for in…
Copilot Jan 30, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 3 additions & 2 deletions backend/app/services/music_service.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -844,8 +844,9 @@ def _pad_audio_token(token):
else:
raise RuntimeError("Cannot load HeartCodec: codec_path not available")

# Convert frames to codec device and dtype (important for MPS float16)
frames_for_codec = frames.to(device=pipeline.codec_device, dtype=codec_dtype)
# Move frames to codec device (keep dtype as long for indexing)
# frames contains token IDs (integers) used as indices, so dtype must remain long
frames_for_codec = frames.to(device=pipeline.codec_device)
wav = pipeline.codec.detokenize(frames_for_codec)

# Cleanup codec if using lazy loading (free VRAM for next generation)
Expand Down