Announcing Ultralytics v8.3.60 Release!
We are thrilled to announce the release of Ultralytics v8.3.60, packed with key updates to enhance the user experience, improve performance, and streamline development workflows. This update includes important improvements across CoreML, Docker environments, auto-annotation guides, and more β all while maintaining smooth compatibility. Read on for the full feature breakdown!
Summary
This release focuses on resolving specific CoreML segmentation output issues, upgrading Docker for faster container workflows, and enriching documentation for segmentation auto-annotation. Other updates like improved bug reporting templates and standardized code formatting further refine the platform for developers and users!
Key Changes
CoreML Segmentation Fix
- Update: Fixed logic in
autobackend.py
for CoreML segmentation outputs to handle reverse order cases correctly. - Impact: Seamless deployment on Apple devices with reduced error risks for segmentation workflows.
Docker Update
- Update: Upgraded Dockerfile to PyTorch 2.5.1 with CUDA 12.4 and cuDNN 9.
- Impact: Enhanced performance and compatibility for containerized environments, ensuring faster training and inference workflows.
Colab Integrations
- Update: Added Colab badges across documentation to simplify model experimentation with interactive setups.
- Impact: New users can jump into hands-on tutorials quickly, easing the learning curve.
Enhanced Auto-Annotation Documentation
- Update: Improved clarity of segmentation auto-annotation guides, covering MobileSAM and other models.
- Impact: Streamlined dataset labeling for large-scale projects, saving setup and implementation time.
Bug Reporting Improvements
- Update: Issue templates now require full traceback data for streamlined debugging.
- Impact: Quicker bug fixes and more efficient responses from the development team.
Standardized String Formatting
- Update: Consolidated codebase to use double-quoted f-strings for better readability and maintainability.
- Impact: Cleaner, more cohesive code for a smoother developer experience.
Purpose & Impact
- CoreML Segmentation: Improved model support for Apple-device-specific workflows, reducing segmentation output errors.
- Docker Upgrade: Modernized container workflows for improved speed and reliability across cloud & local setups.
- Colab Badges: Empowered developers with interactive resources, making experimentation simple and accessible.
- Auto-Annotation: Comprehensive, updated guides reduce friction for dataset labeling tasks.
- Bug Reporting: Enhanced detail collection for faster resolution of user-reported issues.
- Code Cleanliness: Aligned string formatting fosters long-term maintainability in a collaborative codebase.
Importantly, the release introduces no breaking changes, so you can upgrade your workflows seamlessly.
Whatβs Changed
- Apply Ruff 0.9.0 by @glenn-jocher
- Add new Colab Notebooks badges to Docs by @RizwanMunawar
- Update
val.md
by @RizwanMunawar - Update issue templates with better instructions by @Y-T-G
- Upgrade Dockerfile to PyTorch 2.5.1 by @glenn-jocher
- Add warning for
task=classify
withmode=track
by @RizwanMunawar - Add
mobile-sam
auto-annotation to segmentation datasets docs by @RizwanMunawar - Fix CoreML Segment inference by @Y-T-G
Full Changelog: v8.3.60 Release Notes
Try It Out!
We encourage you to upgrade to v8.3.60 and explore these improvements firsthand. Whether youβre deploying CoreML for Apple, using Docker workflows, or experimenting with new Colab guides, this release offers something for everyone!
Download and Upgrade Ultralytics
View the Documentation
Submit Feedback
Thank you for being a part of our growing community! Your feedback fuels our progress and ensures we continue delivering state-of-the-art AI solutions.