Change8

v5.15.0

📦 invokeaiView on GitHub →
5 features🐛 5 fixes1 deprecations🔧 2 symbols

Summary

This release introduces support for FLUX LoRAs and SDXL IP Adapter Plus, alongside significant stability improvements including fixes for rare validation errors and memory usage optimization. Users are encouraged to use the Invoke Launcher for installation and updates.

Migration Steps

  1. The Invoke Launcher is the recommended way to install, update, and run Invoke. Follow the Quick Start guide if using the launcher.
  2. If performing a headless install, follow the manual install guide.

✨ New Features

  • Support for AI Toolkit FLUX LoRAs.
  • SDXL IP Adapter Plus added to starter models.
  • Gallery search now supports filtering by image creation dates.
  • Improved JSON formatting via updated fracturedjsonjs to version 4.1.0.
  • Support for CUDA devices in slots 2 and above.

🐛 Bug Fixes

  • Fixed `AttributeError: module 'cv2.ximgproc' has no attribute 'thinning'` error when using Control Layers.
  • Fixed a rare `ValidationError` occurring during generation/dequeuing (related to #7950).
  • Fixed an error on upload + resize operations for images with unusual modes.
  • Fixed possible race conditions during job enqueueing.
  • Resolved an OpenCV dependency conflict.

🔧 Affected Symbols

cv2.ximgprocfracturedjsonjs

⚡ Deprecations

  • Internal code migrated from using the deprecated `warn` function to the modern `warning` interface.