●Change8

v0.3.50

📦 comfyuiView on GitHub →
✨ 5 features🐛 6 fixes🔧 3 symbols

Summary

This release introduces significant support for Qwen Image models, including merging nodes and LoRA support, alongside various stability fixes and asynchronous API enhancements. It also updates internal templates and testing workflows.

Migration Steps

  1. Update template versions (0.1.52, 0.1.58, 0.1.59).
  2. Update frontend version to v1.24.4 (though later downgraded for release, users might need to check current state).

✨ New Features

  • Added Qwen Image model support to readme.
  • Implemented Qwen Image model merging node.
  • Added async API nodes support.
  • Added support for SimpleTuner lycoris lora for Qwen-Image.
  • Added Wan2.2 fun control support.

🐛 Bug Fixes

  • Fixed issue where qwenLora could not load properly.
  • Fixed RepeatLatentBatch not working on multi-dimensional latents.
  • Added safety check for torchaudio import in _ui.py.
  • Addressed potential crash issues related to gfx1201 on flux with pytorch attention.
  • Ensured AMD compatibility by preventing crashes when fp16 acc is not supported.
  • Fixed Kling Image API Node: stopped passing "image_type" when no image is present.

🔧 Affected Symbols

_ui.pyRepeatLatentBatchKling Image API Node