New Release: Ultralytics v8.3.140

[glenn-jocher]

Ultralytics YOLO v8.3.140 Release — Smoother Installs, Smarter Exports, and Better Learning Tools!

Hi everyone,

We’re excited to announce the release of Ultralytics YOLO v8.3.140! This update focuses on reliability, flexible data exports, expanded hardware support, and even better resources for learning and deploying YOLO models.

---

Summary

This release brings improved reliability for package management, enhanced export and benchmarking features, clearer documentation, and new video tutorials to help users get the most out of Ultralytics YOLO models.

---

New Features & Major Improvements

Improved Package Management

• Smarter “uv” Detection: Installing dependencies is now smoother thanks to a more robust check for the “uv” package manager, significantly reducing setup issues.
  • PR by @glenn-jocher

Flexible Data Exports

• Enhanced to_df and to_sql Methods: Control your export normalization, decimal precision, and create SQL tables dynamically for streamlined analysis and sharing.
  • to_df improvements by @Laughing-q
  • to_sql enhancements by @RizwanMunawar

Expanded Benchmarking & Testing

• SONY IMX Support: Added IMX as a backend for benchmarking, opening new hardware possibilities.
  • Docstring for IMX benchmarks by @lakshanthad
• Jetson ONNX Testing: Re-enabled ONNX export tests on NVIDIA Jetson devices, improving edge support.
  • ONNX Jetson CIs by @lakshanthad

Documentation & Learning

• Reorganized Integrations: Cleaner docs make it easier to find and deploy new integrations.
  • Docs reorganization by @glenn-jocher
• SONY IMX500 Highlighted: See how to run YOLO models on this innovative hardware.
• New Video Tutorials: Learn DOTA dataset training and thread-safe inference directly from embedded YouTube videos.
  • Tutorial integrations by @RizwanMunawar

---

Improvements

• Cleaner Test Suite: Removed unnecessary Python version checks for more streamlined testing and maintenance.
  • Test condition improvements by @RizwanMunawar

---

Purpose & Impact

This update is all about making Ultralytics tools more:

• Reliable: Smoother installations with improved dependency handling.
• Flexible: Advanced data export tools for easier, richer analyses.
• Accessible: See expanded hardware support, especially for Jetson and IMX platforms.
• User-Friendly: More discoverable docs and hands-on video guides.

Whether you’re developing new features, training new models, or simply exploring computer vision, Ultralytics YOLO v8.3.140 offers meaningful improvements for every step of your workflow!

---

How to Get Involved

• Try out the new release: Update now and let us know how it’s working for you.
• Share your feedback: We’d love to hear your questions, suggestions, and successes in Ultralytics Discussions.
• Read the full changelog: For all details, see the complete comparison.
• See the release page: Ultralytics v8.3.140 Release

---

A big thank you to everyone in the YOLO community and the Ultralytics team for your continued contributions and support! Your feedback drives each and every release.

Happy experimenting!

— The Ultralytics Team