Beta Release Notes

YORU v2.0.0-beta.1

Pre-release software. This version targets the develop2 branch and may contain bugs. Feedback and bug reports via GitHub Issues are welcome.

To use the beta version, check out the corresponding tag:

git checkout v2.0.0-beta.1

New Features

Model Support Expansions

• YOLOv8 / YOLO11: Accessible through a unified model wrapper.
• RT-DETR: Real-Time Detection Transformer support added.
• Torchvision models: Faster R-CNN, Mask R-CNN, and SSD variants now available.

Interface Enhancements

• Real-time configuration tool: A new configuration creation tool built with DearPyGui allows interactive setup.
• Training interface improvements:
  • Progress visualization during training.
  • Automatic state recovery after interruption.
  • Layout refinements for better usability.
  • Automatic model detection.

Changes

Code Organization

• Libraries reorganized into a yoru/ package structure for cleaner imports and maintainability.
• The labelImg annotation tool is now bundled directly into YORU with several bug corrections applied.

Bug Fixes

• Resolved PyTorch 2.6 compatibility issues specific to YOLOv5 inference.

Version History