Ultralytics v8.3.165 Release – Unified Dataset Paths for Smoother Workflows!
We’re excited to announce the release of Ultralytics v8.3.165! This update is all about making dataset management more intuitive and consistent for everyone, whether you’re just starting out or are a seasoned YOLO user.
Summary
Ultralytics v8.3.165 brings a major quality-of-life improvement by standardizing the folder structure for datasets. Now, all datasets adopt a unified image directory convention, making configuration easier and reducing confusion in your projects.
New Features
Unified Dataset Structure
- Consistent Path Convention: All image directories in supported YAML configuration files now follow the structure:
images/trainimages/valimages/test
- Streamlined Comments: Updated dataset comments provide clearer notes about download sizes.
Improvements
- Reduced Confusion: With one standardized way to specify image folders, you’ll spend less time troubleshooting path errors and more time training models.
- Simpler Automation: Automate dataset switching and experiment in a snap—no more adjusting YAMLs for each dataset.
- Easier Maintenance: These changes lay the groundwork for clearer documentation, future releases, and a better developer experience overall.
PRs & Contributors
Special thanks to @RizwanMunawar for leading this improvement in PR #21353!
Want to see the full list of updates? Check the release changelog.
Try It Out & Share Your Thoughts
You can find all the details in the v8.3.165 release notes. Update today and let us know how the new unified structure simplifies your workflow. Your feedback is essential—join the discussion and help shape future releases!
As always, this progress reflects the collective effort and dedication of the YOLO community and the Ultralytics team. Thank you for your ongoing support, collaboration, and feedback.
Happy training!