ultralytics 8.3.153

v8.3.153 - `ultralytics 8.3.153` OpenVINO COCO128 benchmarks (#20877)

on Python PyPI

🌟 Summary

This release introduces comprehensive OpenVINO benchmark results for YOLO11 models on the latest Intel® Core™ Ultra™ 7 265K hardware, alongside improved documentation, more accurate benchmarking data for Raspberry Pi and Rockchip devices, enhanced tracking examples, and a key fix for per-class metric reporting. 🚀

📊 Key Changes

• OpenVINO Benchmarks on Intel Core Ultra 7 265K: Added detailed performance tables and charts for YOLO11 (n, s, m, l, x) models across GPU, CPU, and NPU, covering both FP32 and INT8 precisions.
• Raspberry Pi & Rockchip Docs Updated: Refreshed benchmark results and instructions for YOLO11 on Raspberry Pi and Rockchip RKNN, using the COCO128 dataset for more accurate, up-to-date performance data.
• Tracking Example Improvements: Simplified and clarified YOLO11 tracking code and documentation, making it easier to track objects across multiple video streams.
• Per-Class Metrics Fix: Corrected class name alignment in detection, segmentation, and pose metric summaries, ensuring accurate per-class evaluation results.
• Minor Documentation Fixes: Corrected dataset naming and improved clarity in various docs.

🎯 Purpose & Impact

• Empowers Hardware Choices: Users can now make informed decisions about deploying YOLO11 with OpenVINO on the newest Intel hardware, maximizing speed and efficiency for their applications.
• Boosts Transparency & Reproducibility: Updated benchmarks and clearer instructions help users understand expected performance and reproduce results on their own devices, whether using Raspberry Pi, Rockchip, or Intel platforms.
• Easier Multi-Video Tracking: Enhanced tracking examples lower the barrier for implementing robust, scalable object tracking solutions.
• Trustworthy Evaluation: The metrics fix ensures users see correct, clearly labeled per-class results, improving confidence in model evaluation and comparison.
• Overall Smoother Experience: Documentation and usability improvements benefit both beginners and advanced users working with YOLO11 and Ultralytics HUB. 💡

This update is especially valuable for developers and researchers deploying YOLO11 models on modern Intel chips or edge devices, and for anyone seeking reliable, easy-to-understand benchmarking and tracking workflows.

What's Changed

• Update trackers README.md by @RizwanMunawar in #20986
• Update Raspberry Pi Doc with COCO128 Benchmarks by @lakshanthad in #20990
• Update Rockchip RKNN Doc with COCO128 Benchmarks by @lakshanthad in #20991
• Fix COCO128 typo in rockchip-rknn.md by @RizwanMunawar in #21006
• Fix incorrect category index in metrics.summary by @Laughing-q in #20995
• ultralytics 8.3.153 OpenVINO COCO128 benchmarks by @ambitious-octopus in #20877

Full Changelog: v8.3.152...v8.3.153