v8.3.23 - `ultralytics 8.3.23` fix `bbox2segment` when no segments generated (#17157)

🌟 Summary

The v8.3.23 release of Ultralytics YOLO introduces stability improvements by fixing a bug in the data conversion process, updating versioning, and refining user settings and documentation.

📊 Key Changes

• Version Update: The software version is updated from 8.3.22 to 8.3.23.
• Bug Fix in Data Conversion: Ensures yolo_bbox2segment skips generating empty segment lists, preventing potential errors.
• Reduced Python Warnings: Edited Python version check to minimize console spam.
• Documentation Update: Corrected export format examples for INT8 quantization, aligning with TensorRT capabilities.
• W&B Logger Default: Weights & Biases logging is now disabled by default.
• Environment Detection: Enhanced accuracy for identifying Jupyter environments.

🎯 Purpose & Impact

• 🚀 Improved Stability: Fixing the bug in data conversion leads to more reliable performance during bounding box to segment transformations.
• 📉 Cleaner Console: Less console clutter from Python checks makes for a smoother user experience.
• 📝 Clearer Documentation: Updated docs guide users on proper export procedures, easing model deployment tasks.
• 📈 Optimized Resource Use: Disabling Weights & Biases by default reduces unnecessary compute and network usage unless needed.
• 🧠 Reliable Environment Behavior: Accurate environment detection prevents misidentification in diverse setups, adapting better to where the software runs.

What's Changed

• Fix Python warning spam by @Y-T-G in #17162
• Fix inaccurate example in Export docs by @Y-T-G in #17161
• Default W&B setting False by @glenn-jocher in #17164
• ultralytics 8.3.23 fix bbox2segment when no segments generated by @Laughing-q in #17157

Full Changelog: v8.3.22...v8.3.23

v8.3.22 - `ultralytics 8.3.22` SAM2.1 integration (#17131)

🌟 Summary

The Ultralytics v8.3.22 release brings significant improvements by integrating the SAM 2.1 model, enhancing segmentation capabilities, and implementing various fixes and updates for better user experience and functionality.

📊 Key Changes

• SAM 2.1 Integration: Introduces support for the SAM 2.1 model across all scales, upgrading pre-trained weights to version 8.3.0.
• Device Handling Fix: Improves logic for exporting models to TensorRT, ensuring correct device processing.
• Configuration Updates: Streamlines solution-specific default configurations directly within the code.
• Binder Integration: Adds a Binder badge to allow running Ultralytics in an interactive Jupyter notebook environment.

🎯 Purpose & Impact

• Improved Segmentation: SAM 2.1 enhances segmentation accuracy with advanced algorithms and features like spatial memory handling and temporal encoding, benefiting users needing precise object segmentation. 🎨
• Robust Exporting: Fixes in device handling bolster reliability when exporting models, preventing potential errors and aiding in smoother operation. ⚙️
• User Experience Enhancement: The consolidation of configuration management reduces complexity, providing a more seamless user setup process. 🛠️
• Accessibility: The Binder integration makes Ultralytics more accessible, allowing users to easily experiment with the software in a flexible, online environment. 🌐

What's Changed

• Fix DLA export when device=None by @Laughing-q in #17128
• Enable default cfg for similar args in solutions. by @RizwanMunawar in #17112
• Add Binder Notebook badge by @glenn-jocher in #17074
• ultralytics 8.3.22 SAM2.1 integration by @Laughing-q in #17131

Full Changelog: v8.3.21...v8.3.22

v8.3.21 - `ultralytics 8.3.21` NVIDIA DLA export support (#16449)

🌟 Summary

The 'v8.3.21' release for Ultralytics introduces NVIDIA Deep Learning Accelerator (DLA) export support, alongside several documentation and usability enhancements.

📊 Key Changes

• NVIDIA DLA Support: Added export options for NVIDIA DLA on Jetson devices, allowing more energy-efficient model inference.
• Documentation Updates: Clarified guides on using TensorRT and DLA for Jetson devices. Updated image URLs in documentation for consistency.
• Comet Integration: Improved logging of plots and metrics for better tracking in Comet during model training and evaluation.
• Parameter Additions: Introduced project and name parameters to better organize prediction and validation outputs.
• Dataset Naming Standardization: Changed "Roboflow 100" to "RF100" for documentation precision.

🎯 Purpose & Impact

• Energy Efficiency: Utilizing NVIDIA DLA can significantly reduce power consumption during inference on Jetson devices, though with some added latency. This is ideal for energy-conscious applications. ⚡
• Enhanced User Experience: Updated documentation provides clearer, more accessible information, and new parameters help in managing results more effectively. 📘
• Improved Metric Tracking: Expanded Comet integration enhances the ability to monitor and analyze different metrics, supporting comprehensive model evaluation. 📊
• Consistency and Reliability: Documentation improvements ensure that links and names are up-to-date, preventing broken references and confusion. 🔗

What's Changed

• Update YOLO11 comparison plots by @glenn-jocher in #17097
• Add YOLO publication notice in Docs by @RizwanMunawar in #17095
• Add project and name args to docs for predict and val task by @Y-T-G in #17114
• Comet integration fix by @yaricom in #17099
• Update datasets index.md by @glenn-jocher in #17098
• ultralytics 8.3.21 NVIDIA DLA export support by @justincdavis in #16449

New Contributors

• @justincdavis made their first contribution in #16449
• @yaricom made their first contribution in #17099

Full Changelog: v8.3.20...v8.3.21

v8.3.20 - `ultralytics 8.3.20` W&B `plots=False` logging fix (#17088)

🌟 Summary

The release of Ultralytics version 8.3.20 introduces improvements for efficiency and usability in both training logging and Docker image compatibility.

📊 Key Changes

• W&B Integration Fix: Adjusted the Weights & Biases logging to prevent errors when plots are deliberately disabled by the user.
• Docker Update: Updated the base Docker image to use a more recent version of PyTorch.
• Pretrained Model Documentation: Added examples for using pretrained YOLO models with Open Images Dataset V7.

🎯 Purpose & Impact

• Improvement in Efficiency: The W&B logging fix optimizes the training process by avoiding unnecessary plotting operations, thereby saving computational resources. 📉
• Enhanced Compatibility: By updating the Docker's PyTorch version, users benefit from potential performance boosts and better support for current CUDA features, facilitating more efficient processing. 🚀
• Better Usability: The addition of code examples for pretrained models makes it easier for users to implement sophisticated AI functionality without a steep learning curve, boosting productivity and accessibility in AI projects. 🧑‍💻

What's Changed

• Dockerfile FROM pytorch/pytorch:2.5.0-cuda12.4-cudnn9-runtime by @glenn-jocher in #17094
• Add Open Images Dataset V7 pretrained model usage examples by @Y-T-G in #17090
• ultralytics 8.3.20 W&B plots=False logging fix by @Anzhc in #17088

New Contributors

• @Anzhc made their first contribution in #17088

Full Changelog: v8.3.19...v8.3.20

v8.3.19 - `ultralytics 8.3.19` TensorRT 10.5.0 support (#17076)

🌟 Summary

The release of ultralytics v8.3.19 provides important updates to the TensorRT support, removing specific versions that caused issues to enhance stability and compatibility.

📊 Key Changes

• TensorRT Updates: Removed the version pinning for tensorrt-cu12 from version 10.1.0, which was previously causing issues.
• Documentation Improvements: Enhanced the script for building documentation with better URL handling and added the ability to trigger documentation publishing manually.

🎯 Purpose & Impact

• Stability Enhancements: By excluding TensorRT version 10.1.0, where known issues were present, this release aims to make model exports more reliable, particularly when using TensorRT.
• Broader Compatibility: Allowing compatibility with more versions of TensorRT can lead to reduced installation issues and more flexible deployment across different systems.
• User Experience: Improved handling of links in documentation ensures that users can navigate resources easily, and the added language support in themes expands accessibility.
• Simplified Code Examples: Adjustments in notebook examples make it easier for developers to perform tasks like heatmap generation and object counting using the YOLO model, streamlining initial setups.

These updates cater to both developers who require stable deployments and non-expert users who benefit from improved documentation and example simplifications. Each change reflects Ultralytics' commitment to improving performance and user experience. 🚀

What's Changed

• Fix build_docs regex for trailing URL periods by @pderrenger in #17036
• Fix MNIST link by @pderrenger in #17038
• Improve Docs publish.yml workflow dispatch by @glenn-jocher in #17039
• Update extra.js by @glenn-jocher in #17040
• Fix Giscus load after navigation by @glenn-jocher in #17042
• Remove YOLO11n message from AMP logs by @Y-T-G in #17075
• Update notebooks by @RizwanMunawar in #17065
• ultralytics 8.3.19 TensorRT 10.5.0 support by @glenn-jocher in #17076

Full Changelog: v8.3.18...v8.3.19

v8.3.18 - Ultralytics Refactor https://ultralytics.com/actions (#17031)

🌟 Summary

This release, version 8.3.18, brings improvements to make navigating documentation easier and enhances user control over logging features.

📊 Key Changes

• Enhanced the conversion of URLs in documentation to clickable links, removing reliance on the BeautifulSoup library.
• Simplified the YOLO HUB login command by removing the now redundant "hub" keyword.
• Changed default settings to disable Weights & Biases logging integration, enhancing user privacy.
• Corrected previous updates that inadvertently broke documentation links by reversing specific changes.

🎯 Purpose & Impact

• Improved Usability: Making URLs clickable in documentation enhances navigation, making it easier for users to explore related content.
• Streamlined Workflow: Simplifying the login process helps users quickly access the Ultralytics HUB with a straightforward command.
• Increased Privacy Control: Disabling automatic logging to Weights & Biases by default respects user preferences regarding third-party data sharing.
• Documentation Integrity: Reverting previous changes ensures that all documentation links remain functional, providing consistent access to resources.

These updates are focused on enhancing user experience and improving code efficiency in documentation handling.

What's Changed

• Update YOLO login by @glenn-jocher in #17022
• ultralytics 8.3.18 fix is_jupyter() to globals() by @glenn-jocher in #17023
• Revert "Fixed build docs regex security" by @UltralyticsAssistant in #17027
• Ultralytics Refactor https://ultralytics.com/actions by @UltralyticsAssistant in #17031

Full Changelog: v8.3.17...v8.3.18

v8.3.17 - `ultralytics 8.3.17` accept spaces in CLI args (#16641)

🌟 Summary

The latest release, v8.3.17, enhances command-line argument handling and boosts backward compatibility for older YOLO model versions.

📊 Key Changes

• Command Line Update: Improved how command-line arguments are processed, especially when using spaces and special characters like brackets.
• Backward Compatibility: Reintroduced support for legacy models (v3, v5, v8, v9) to ensure they work smoothly with current updates.

🎯 Purpose & Impact

• Better CLI Experience: The command-line interface now handles complex argument patterns more effectively, making it easier for users and scripts to interact without errors. This is particularly useful for advanced users who configure models via scripts. 🛠️
• Legacy Model Support: Users of older YOLO models can now integrate them with new updates without needing to upgrade models immediately, facilitating a more seamless transition. This helps maintain consistency in performance and saves time on reconfiguration. 📈

By refining command-line interactions and supporting older models, this update aims to improve user convenience and software flexibility.

What's Changed

• Backward compatibility support for legacy models by @Laughing-q in #17010
• ultralytics 8.3.17 accept spaces in CLI args by @glenn-jocher in #16641

Full Changelog: v8.3.16...v8.3.17

v8.3.16 - `ultralytics 8.3.16` PyTorch 2.5.0 support (#16998)

🌟 Summary

The ultralytics v8.3.16 update is a minor release focused on improvements to documentation readability, support for the latest PyTorch version, and minor bug fixes.

📊 Key Changes

• PyTorch 2.5.0 Support: Updates facilitate compatibility with the latest PyTorch version, ensuring users can leverage its enhancements and features.
• Documentation Enhancements: README and other documents were updated to improve layout, navigation, link accuracy, and partner information, enhancing clarity and accessibility.
• Parking Management UI: The Tkinter user interface for the parking management solution was refined to improve usability and code maintainability.
• Docker Image Publishing: Adjustments made to improve Docker image handling for applications using Tkinter.
• Regex Security Update: Improved regex patterns in documentation to enhance link formatting security.

🎯 Purpose & Impact

• Broad Compatibility: Supporting PyTorch 2.5.0 allows users to benefit from performance improvements and new capabilities in their workflows.
• Enhanced User Experience: Documentation improvements make it easier for users to understand and navigate resources, which can aid in learning and implementing the Ultralytics solutions.
• Improved UI & Functionality: The parking management solution's UI improvements will make it more intuitive for users to set up and manage parking zones.
• Security and Optimization: Code refactoring, Docker handling improvements, and enhanced regex contribute to better security, efficiency, and maintainability.

These changes collectively aim to enhance user experience, keep the project updated with industry standards, and ensure robust, secure deployments.

What's Changed

• Update parking-management solution by @RizwanMunawar in #16990
• Update Integrations table by @glenn-jocher in #16994
• Docker tkinter fix by @RizwanMunawar in #16995
• Update README links by @glenn-jocher in #16996
• Fixed build docs regex security by @ambitious-octopus in #17012
• ultralytics 8.3.16 PyTorch 2.5.0 support by @glenn-jocher in #16998

Full Changelog: v8.3.15...v8.3.16

v8.3.15 - `ultralytics 8.3.15` new TPU device-selection ability (#16576)

🌟 Summary

The v8.3.15 release of Ultralytics introduces enhanced support for selecting TPU devices, improved code readability, and refines several workflows and documentation aspects.

📊 Key Changes

• TPU Device Selection: Added functionality to choose specific TPU devices when multiple are available, essential for managing resources across multiple containers.
• Code Refactoring: Improved code clarity and maintainability in autobackend.py.
• Version Management: Simplified PyPI version checking logic for more efficient publishing workflows.
• Documentation Updates:
  • Expanded CI tables to include more repositories and added clarity to link formatting.
  • Corrected documentation URLs and clarified usage instructions for TPU models.

🎯 Purpose & Impact

• Enhanced Flexibility: Users can now run applications on specific TPUs, preventing conflicts and optimizing resource allocation — especially useful for complex, multi-container setups. ⚙️
• Code Maintenance: The refactoring improves readability and ease of future maintenance, which is beneficial for all developers engaging with the project. 🛠️
• Streamlined Release Process: By minimizing the complexity of the version management script, the stability and efficiency of release cycles are improved, reducing the potential for human error. 🚀
• Improved User Guidance: More intuitive and accurate documentation means users will experience fewer issues and a better understanding of product capabilities. 📚

This release makes using multiple TPU resources smoother and improves the overall quality and maintainability of the project infrastructure.

What's Changed

• Update publish.yml by @glenn-jocher in #16967
• Expand Docs CI table with mkdocs, thop, actions, handbook by @glenn-jocher in #16970
• Fix broken Tencent AI link by @glenn-jocher in #16971
• Remove onnxslim==0.1.34 pin by @glenn-jocher in #16974
• Update utralytics/solutions docstrings to match new YOLO11 by @jk4e in #16975
• Fix display of links in yolov5 docs by @jk4e in #16973
• Fix Triton inference without explicit metadata by @zeograd in #16938
• ultralytics 8.3.15 new TPU device-selection ability by @lalayants in #16576

New Contributors

• @lalayants made their first contribution in #16576
• @zeograd made their first contribution in #16938

Full Changelog: v8.3.14...v8.3.15

v8.3.14 - `ultralytics 8.3.14` update TensorRT `dynamic` inference (#16953)

🌟 Summary

The release of v8.3.14 brings significant improvements to dynamic inference handling in the TensorRT engine and several documentation updates.

📊 Key Changes

• TensorRT Inference Update: Improved condition handling for image shape management in TensorRT's autobackend module.
• Documentation Corrections: Fixed links, naming errors, and clarified documentation for various user guides and integration instructions.
• Code Refactor: Optimized the codebase for better performance and readability.

🎯 Purpose & Impact

• Enhanced Performance: The TensorRT update optimizes image shape checking, which boosts performance and reduces errors during dynamic input handling.
• Improved User Guidance: Documentation enhancements provide clearer and more accurate guidance, helping both new and experienced users to navigate and utilize Ultralytics features effectively.
• Streamlined Codebase: Refactoring leads to quicker code execution and simplifies maintenance, benefiting developers working with the project.

These updates collectively aim to enhance the efficiency and user experience of the Ultralytics platform. 🌟🚀

What's Changed

• Fix link in docs guide page by @jk4e in #16932
• Fix naming of dataset for OBB in Google Colab Tutorial by @jk4e in #16937
• Ultralytics Code Refactor https://ultralytics.com/actions by @UltralyticsAssistant in #16940
• Comet ML Docs - Fix link, naming and formatting by @jk4e in #16941
• Improve Home "Where to Start" Layout in Docs by @jk4e in #16846
• Recommend newer Python version in conda docs by @Y-T-G in #16946
• Update publish.yml by @glenn-jocher in #16948
• Fix Home "Where to Start" grid card link alignment in docs by @jk4e in #16945
• Optimize Example YOLO post-processing speed by @franklinOliveira in #16821
• Update CLA.md by @glenn-jocher in #16913
• Fix self.type variable name in Analytics solution by @RizwanMunawar in #16965
• ultralytics 8.3.14 update TensorRT dynamic inference by @Laughing-q in #16953

New Contributors

• @franklinOliveira made their first contribution in #16821

Full Changelog: v8.3.13...v8.3.14