v0.3.33
Breaking Changes📦 comfyuiView on GitHub →
⚠ 1 breaking✨ 3 features🐛 5 fixes🔧 5 symbols
Summary
This release introduces initial support for the ACE-Step model, improves memory estimation for LTXV models, and resolves several bugs related to ACE functionality and character encoding. Torchaudio is no longer a hard dependency.
⚠️ Breaking Changes
- Loading of api nodes will fail if the installed pyav version is too old. Users must update pyav to a compatible version to continue using api nodes.
Migration Steps
- Ensure your pyav installation is up-to-date to prevent failures when loading api nodes.
✨ New Features
- Implemented initial ACE-Step model support.
- Added experimental lyrics strength control for ACE features.
- Improved detection logic to allow ltxv models without configuration files to load.
🐛 Bug Fixes
- Fixed an issue where chroma was using an incorrect shift; now uses Flux shift.
- Improved memory estimation specifically for the LTXV 13B model.
- Adjusted memory estimation code to correctly handle ACE VAE.
- Ensured Japanese hiragana and katakana characters function correctly with ACE.
- Fixed an issue preventing ACE VAE tiling from working.
🔧 Affected Symbols
api nodeschromaltxv modelsACE-Step modelACE VAE