Ultralytics

🌟 Summary
The `v8.2.85` release for Ultralytics YOLO highlights the latest updates, focusing on the upcoming YOLOv10 enhancements, streamlined workflows, and improved compatibility across the platform.

📊 Key Changes
- **YOLOv10 Parameter Fix**: Introduced a `max_det` parameter, enhancing user control by allowing the specification of the maximum number of detections. This update is a pivotal improvement leading up to YOLOv10, aligning with ongoing efforts to optimize model configurability and performance.
- **GitHub Actions Update**: Streamlined the publish workflow by removing the `openai` dependency and consolidating complex scripts into a single, efficient command. This significantly simplifies the release process.
- **INT8 Export Warning**: Added a warning for automatic enforcement of `dynamic=True` during INT8 model exports to enhance compatibility, ensuring smoother user experiences with advanced export settings.
- **Documentation Enhancements**: Improved author attribution with avatars in documentation, optimizing related tools and making contributions more visible and accessible to users.
- **Explorer Tests Requirement**: Updated tests to require PyTorch version 1.13 or newer, ensuring compatibility, reliability, and stability across development environments.

🎯 Purpose & Impact
- **YOLOv10 Readiness**: The new `max_det` parameter is a key enhancement that sets the stage for YOLOv10, giving users greater control and customization options over model outputs.
- **Optimized Release Workflow**: The streamlined workflow reduces complexity and potential dependency issues, facilitating faster and more efficient publishing of updates.
- **Enhanced Export Compatibility**: The INT8 export warning for `dynamic=True` ensures compliance with optimal settings, improving export reliability and performance.
- **Improved Documentation**: Enhanced visualization with author avatars increases transparency and improves user interaction with the documentation.
- **Reliable Testing**: By enforcing minimum version requirements for PyTorch, the update guarantees stable and consistent testing, enhancing overall development reliability.

These updates emphasize Ultralytics' commitment to enhancing YOLOv10's functionality, improving user control, and refining the overall development and deployment experience.

What's Changed
* Optimize docs author avatars by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15923
* Explorer tests require torch>=1.13 by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15930
* Update notebooks: Fix `classes_names` argument with `names` by RizwanMunawar in https://github.com/ultralytics/ultralytics/pull/15915
* Includes warning for enforced `dynamic` during INT8 exports by Burhan-Q in https://github.com/ultralytics/ultralytics/pull/15874
* Update publish.yml by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15932
* Update publish.yml by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15933
* `ultralytics 8.2.85` YOLOv10 `max_det` arg fix by ambitious-octopus in https://github.com/ultralytics/ultralytics/pull/15917


**Full Changelog**: https://github.com/ultralytics/ultralytics/compare/v8.2.84...v8.2.85

🌟 Summary
This update introduces flexible image size inference for SAM2 using the `ultralytics` package, alongside improvements to testing and documentation.

📊 Key Changes
- **Flexible SAM2 Image Size Inference:** SAM2 now supports custom image sizes through the `ultralytics` package, allowing inference at sizes like 640x640 instead of the default 1024x1024.
- **Testing Enhancement:** Updated CI testing workflow for version-specific compatibility.
- **Documentation Update:** Refreshed documentation with higher quality images.

🎯 Purpose & Impact
- **Enhanced Flexibility:** Users can now run SAM2 inference using the `ultralytics` package with custom `imgsz` values (e.g., `imgsz=640`), offering significant advantages:
- Faster processing times for smaller images
- Reduced memory usage, enabling inference on devices with limited resources
- Maintains good segmentation quality while allowing size-performance tradeoffs
- **Improved Efficiency:** The ability to use smaller image sizes (like 640x640) instead of the default 1024x1024 can lead to faster inference without significant loss in accuracy for many use cases.
- **Broader Accessibility:** This update makes SAM2 more accessible for users with varying computational resources, as they can now adjust the image size based on their specific needs and hardware constraints.

💻 Usage Example
python
from ultralytics import SAM

Initialize SAM model
model = SAM('sam2_b.pt')

Run inference with custom image size
results = model('path/to/image.jpg', imgsz=640)


This update significantly enhances SAM2's versatility within the `ultralytics` ecosystem, allowing users to fine-tune the balance between speed and accuracy based on their specific requirements.

What's Changed
* Add retry step to failed Conda tests by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15884
* Use AVIF banner images by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15876
* Remove image "?" args by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15891
* fix HUB download and train by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15896
* Optimize Docs images by RizwanMunawar in https://github.com/ultralytics/ultralytics/pull/15900
* Run Conda tests with aligned tag/version by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15904
* Adding missing datasets information to docs by jk4e in https://github.com/ultralytics/ultralytics/pull/15916
* `ultralytics 8.2.84` new SAM flexible `imgsz` inference by Laughing-q in https://github.com/ultralytics/ultralytics/pull/15882


**Full Changelog**: https://github.com/ultralytics/ultralytics/compare/v8.2.83...v8.2.84

🌟 Summary
The v8.2.83 release of Ultralytics introduces improved disk cleanup processes, enhances community engagement options, and makes changes to work with GitHub Actions and documentation.

📊 Key Changes
- **Issue Template Enhancements**: Added Reddit as a new contact link, providing an additional platform for community interaction.
- **Docker Workflow Update**: Improved disk cleanup using a new method to free up to 30GB of space, making builds more efficient.
- **Documentation & Guidance Tweaks**: Updated messages and documentation references to streamline contributions and improve user experience.
- **GitHub Actions Improvements**: Enhanced branch management automation by labeling popular pull requests and clarifying process outputs.
- **VS Code Integration**: Introduced a new VS Code extension with code snippets to accelerate development workflows with Ultralytics.

🎯 Purpose & Impact
- **Community Engagement**: Adding Reddit as a contact method broadens the engagement avenues, fostering a richer community interaction.
- **Efficiency in CI/CD**: The new disk cleanup approach significantly reduces disk usage during CI/CD, accelerating build and release processes.
- **Enhanced User Experience**: Refined documentation and automated pull request updates ensure smoother user and contributor interactions.
- **Developer Productivity**: The VS Code extension aids developers by providing quick access to code snippets, reducing development time and potential errors.

What's Changed
* Update merge-main-into-prs.yml by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15801
* Fix very large banner SVGs bug by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15803
* Rate limit auto-merge action by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15802
* Auto-label PR if above threshold for positive reactions by Burhan-Q in https://github.com/ultralytics/ultralytics/pull/13605
* Add Reddit link to documentation and templates by Y-T-G in https://github.com/ultralytics/ultralytics/pull/15041
* Add https://www.reddit.com/r/Ultralytics/ badge by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15818
* Add VS Code Extension by Burhan-Q in https://github.com/ultralytics/ultralytics/pull/15027
* Cleanup GitHub Runner for Docker CI by lakshanthad in https://github.com/ultralytics/ultralytics/pull/15848
* Fix typo in docstring in `metrics.py` by jk4e in https://github.com/ultralytics/ultralytics/pull/15842
* Add warning message on `distance-calculation.md` docs page by RizwanMunawar in https://github.com/ultralytics/ultralytics/pull/15838
* Fix link in README.zh-CN.md by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15858
* Fix loaders.py broken YouTube example URLs by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15863
* Improve Docs layout issues by jk4e in https://github.com/ultralytics/ultralytics/pull/15868
* Add https://youtu.be/WwXnljc7ZUM to docs by RizwanMunawar in https://github.com/ultralytics/ultralytics/pull/15870
* `ultralytics 8.2.83` fix SAM2 CLI usage by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15871


**Full Changelog**: https://github.com/ultralytics/ultralytics/compare/v8.2.82...v8.2.83

🌟 Summary
The Ultralytics v8.2.82 release introduces significant advancements with the addition of **YOLOv10 export support** for CoreML, Edge TPU, and TF.js, marking this as the standout feature of the update.

📊 Key Changes
- **YOLOv10 Export Support**: Expanded export capabilities now include CoreML, Edge TPU, and TF.js, enabling broader deployment options for YOLOv10 models across different platforms and devices.
- **Docstring Style Adjustments**: Updates to the documentation build workflow to ignore specific style rules for docstrings to align with Google-style conventions.
- **Automated Code Style Checks**: Integration of `ruff` as a new tool to perform style checks and fixes across the codebase, with a particular focus on improving docstring consistency.
- **Examples & Code Cleanup**: Improvements and refinements in example scripts and code snippets, enhancing clarity and consistency, particularly in docstring formats and argument examples.
- **Multilingual Documentation**: Minor updates to documentation links and the addition of new language support for specific parts of the guides.

🎯 Purpose & Impact
- **Enhanced Deployment Options**: The new export capabilities for YOLOv10 models significantly broaden the range of deployment scenarios, making it easier for developers to deploy models on various platforms, including Apple devices, Google's Edge TPU, and web environments via TF.js.
- **Improved Developer Experience**: Consistent and clear docstring styles improve code readability and maintainability, helping developers understand and contribute more efficiently.
- **User-Friendly Examples**: More accurate and consistent example scripts enable users to replicate and learn from them more effectively, reducing errors during implementation.
- **Global Accessibility**: The expansion of multilingual documentation allows a wider global audience to access resources in their preferred languages, enhancing usability and outreach.

What's Changed
* Update robots.txt by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15789
* Properly use cmake variable in ONNXRuntime by memorylorry in https://github.com/ultralytics/ultralytics/pull/15776
* Missing best.pt resumed checkpoint upload spelling by sergiuwaxmann in https://github.com/ultralytics/ultralytics/pull/15772
* 'best.pt' inherit all-epochs results curves from 'last.pt' by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15791
* Ruff format docstring Python code by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15792
* Ruff Docstring formatting by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15793
* TQDM, SimpleClass, IterableSimpleNamespace docstrings by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15795
* Fix YOLOv8 C++ ONNXRuntime transpose op by memorylorry in https://github.com/ultralytics/ultralytics/pull/15779
* `ultralytics 8.2.82` YOLOv10 CoreML, Edge TPU, and TF.js export support by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15796

New Contributors
* memorylorry made their first contribution in https://github.com/ultralytics/ultralytics/pull/15776

**Full Changelog**: https://github.com/ultralytics/ultralytics/compare/v8.2.81...v8.2.82

🌟 Summary
The v8.2.81 release of Ultralytics YOLO introduces significant documentation improvements and a refined model upload process to enhance usability and reliability for machine learning practitioners.

📊 Key Changes
- **Documentation Enhancements**: Updated how code examples and citations are presented across various dataset guides to improve readability and user accessibility.
- **Model Upload Process**: Improved handling of model uploads with added safeguards and informative logging, particularly in transient environments like Google Colab.
- **Technical Adjustment**: Corrected bounding box conversion details in the results module, changing from top-left coordinates to center-point coordinates.

🎯 Purpose & Impact
- **Easier Understanding and Use**: The new documentation format with examples in both Python and CLI makes it more straightforward for users to implement various datasets in their projects.
- **Robust Model Management**: Enhanced model upload functionality ensures continuity and mitigates issues in cloud-based environments, promoting more reliable training sessions.
- **Accurate Representation**: The bounding box conversion adjustment aligns technical descriptions with actual functionalities, ensuring users correctly apply transformations.

What's Changed
* Add https://youtu.be/WwXnljc7ZUM to docs by RizwanMunawar in https://github.com/ultralytics/ultralytics/pull/15751
* Improve Docs dataset layout issues by jk4e in https://github.com/ultralytics/ultralytics/pull/15696
* Fix `xywh` docstring by Laughing-q in https://github.com/ultralytics/ultralytics/pull/15765
* Ensure matplotlib backend gets reset with plt_settings by rschroll in https://github.com/ultralytics/ultralytics/pull/15759
* `ultralytics 8.2.81` fix HUB missing 'best.pt' resumed checkpoint upload by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15754

New Contributors
* rschroll made their first contribution in https://github.com/ultralytics/ultralytics/pull/15759

**Full Changelog**: https://github.com/ultralytics/ultralytics/compare/v8.2.80...v8.2.81

🌟 Summary

The v8.2.80 release of the Ultralytics YOLO software brings a mix of behind-the-scenes improvements, documentation updates, and usability enhancements, particularly around data handling and workflow automation. 📦

📊 Key Changes

- **GitHub Workflows Adjustment**: Modifications made to GitHub Actions workflows to streamline pull request and publication processes by refining the way PR details are extracted.
- **Enhanced Dataset Management**: Added a new section in documentation about downloading datasets directly from Ultralytics HUB.
- **Documentation Language Links Update**: Some corrections and improvements in language link references in both the main README and tutorials.
- **Model Interface Improvement**: Expanded support for input types, including PIL (Python Imaging Library) images, for more versatile image processing.
- **Metrics Code Update**: Tweaked how class names are handled in plotting functions, shifting from list to dictionary format for better identification and flexibility.

🎯 Purpose & Impact

- **Streamlined Processes**: The workflow updates aim to improve developer efficiency and reduce errors in managing code changes, enhancing the overall continuous integration experience. 🚀
- **User-Friendly Dataset Access**: The added documentation helps users easily download datasets, making data preparation more straightforward and less time-consuming. 📚
- **Better Compatibility**: By accepting PIL images, the model becomes more versatile, especially beneficial for users integrating with various image processing libraries. 🖼️
- **Improved Accuracy**: Using dictionaries for class names in metrics promotes better clarity and precision in data interpretation and result visualization, aiding in more insightful analysis. 📈

What's Changed
* Remove Hindi and Dutch languages by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15703
* Remove Hindi and Dutch languages by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15704
* Update publish.yml to use `gh` by glenn-jocher in https://github.com/ultralytics/ultralytics/pull/15686
* Update HUB Docs for new dataset download feature by sergiuwaxmann in https://github.com/ultralytics/ultralytics/pull/15728
* Fix model `names` docstring type to `dict` by yuefanhao in https://github.com/ultralytics/ultralytics/pull/15726
* `ultralytics 8.2.80` add PIL type hints to `model()` by JoshJarabek7 in https://github.com/ultralytics/ultralytics/pull/15719

New Contributors
* yuefanhao made their first contribution in https://github.com/ultralytics/ultralytics/pull/15726
* JoshJarabek7 made their first contribution in https://github.com/ultralytics/ultralytics/pull/15719

**Full Changelog**: https://github.com/ultralytics/ultralytics/compare/v8.2.79...v8.2.80