pypi ultralytics 8.3.169
v8.3.169 - `ultralytics 8.3.169` Add GT, TP, FP, FN visualization for Detect, Segment, Pose and OBB (#18868)

latest releases: 8.3.191, 8.3.190, 8.3.189...
one month ago

🌟 Summary

This release introduces powerful new visualizations for model evaluation, making it easier to understand and debug your results across Detect, Segment, Pose, and Oriented Bounding Box (OBB) tasks. 🖼️✨

📊 Key Changes

  • Ground Truth & Prediction Visualization: You can now visualize ground truths (GT), true positives (TP), false positives (FP), and false negatives (FN) directly from the confusion matrix during validation for Detect, Segment, Pose, and OBB models.
  • New visualize Option: Added a visualize=True argument for validation commands, enabling these detailed visual breakdowns.
  • Documentation Improvements:
    • Embedded a new YouTube tutorial for the HomeObjects-3K dataset, showing how to train and export with YOLO11.
    • Updated CLI and export documentation for clarity and accuracy.
  • Model & Export Enhancements:
    • Improved class name handling for exported models, ensuring safer and more reliable workflows.
    • RT-DETR predictions are now sorted by confidence and limited to the top results for consistency.
    • Simplified TensorFlow export by removing an unnecessary optimization flag.

🎯 Purpose & Impact

  • Easier Debugging & Model Interpretation: The new visualizations help users quickly identify where their models are performing well or struggling, making it much easier to debug and improve results.
  • Better Learning Experience: The added video tutorial and clearer documentation make it simpler for users of all levels to get started and succeed with Ultralytics models.
  • Improved Reliability: Updates to model export and prediction handling reduce errors and make workflows smoother, especially for advanced users working with custom or exported models.
  • Consistent Outputs: RT-DETR and other models now provide more predictable and relevant results, enhancing user confidence in deployment scenarios.

Overall, this update brings valuable tools for both beginners and experts to better understand, visualize, and improve their computer vision projects with Ultralytics. 🚀

What's Changed

New Contributors

Full Changelog: v8.3.168...v8.3.169

Don't miss a new ultralytics release

NewReleases is sending notifications on new releases.