Ultralytics Files

🌟 Summary

Version 8.4.136 improves hyperparameter tuning, inference performance, backend compatibility, and data handling—making YOLO workflows more reliable and efficient. 🚀

📊 Key Changes

• 🎯 Smarter hyperparameter tuning — Current PR [#25996] by @glenn-jocher
• Keeps the existing Gaussian search behavior unchanged during the first 30 completed trials.
• Learns relationships between promising hyperparameters by analyzing the best-performing results.
• Uses confidence-weighted, correlated mutations only when enough elite trial data is available.
• Reflects correlated proposals at search boundaries to avoid repeatedly clipping values.
• Benchmarking on basketball-hoop detection reported a new best fitness of 0.60576, outperforming the tested Ray Tune and previous custom Tuner configurations.

• Updated the hyperparameter tuning documentation.

• 🧠 More robust channels-last inference

• AutoBackend now owns memory-layout selection during backend construction, avoiding duplicated or unsafe conversions.
• Automatic channels-last selection is available for supported Linux and Windows x86 CPU environments using PyTorch 1.13 or newer with oneDNN.
• CUDA support remains available, while ARM64, MPS, older PyTorch versions, and exported backends retain their existing behavior.

• Fixed compatibility issues affecting PyTorch 1.9 and JetPack 6 systems.

• ⚡ Faster image preprocessing

• PIL and NumPy image inputs now use more efficient OpenCV color conversions.

• Avoids unnecessary image copies while preserving correct channel order and contiguous memory layout.

• 🔍 More reliable prediction filtering

• Fixed the CLI classes filter for YOLOE and World models when class IDs are supplied numerically.

• Text-based class prompts continue to work as before.

• 📷 Improved image and visualization handling

• TIFF loading now respects uppercase extensions and grayscale flags, preserving multispectral image channels correctly.
• Pose visualization now scales keypoint coordinates without incorrectly scaling confidence values.

• Matplotlib backend restoration is safer when the originally configured backend is unavailable.

• 🏃 Tracking and distributed tuning improvements

• BoT-SORT sparse optical-flow tracking avoids an unnecessary per-pixel grid allocation, reducing overhead on large frames.

• MongoDB-based distributed tuning now assigns iteration IDs atomically, preventing duplicate trial numbers from concurrent workers.

• 🧪 Better validation and project maintenance

• Dataset YAML files now receive early type validation with clearer error messages.
• Added CI and PyPI publishing status for the Ultralytics SDK repository.
• Updated the package version to 8.4.136.

🎯 Purpose & Impact

• Better tuning results: The Tuner can discover useful relationships between hyperparameters instead of treating every parameter independently, potentially improving final model quality with fewer wasted trials. 📈
• Safer inference across platforms: Backend construction now handles memory formats and retained tensors more consistently, reducing regressions on older PyTorch versions, ARM64 devices, and JetPack environments.
• Faster predictions: Common PIL and NumPy input paths require fewer copies and more efficient conversions, which can improve throughput in image-heavy applications.
• More predictable CLI behavior: Class filtering now works consistently across standard, YOLOE, and World models.
• Improved dataset reliability: Invalid YAML field types are reported earlier, making dataset configuration errors easier to diagnose.
• No architecture changes: This release does not introduce a new model architecture; its primary benefits are improved tuning, compatibility, performance, and correctness.

What's Changed

• Add Platform SDK CI status by @glenn-jocher in https://github.com/ultralytics/ultralytics/pull/25997
• Fix legacy torch predict regressions from default CPU channels-last by @onuralpszr in https://github.com/ultralytics/ultralytics/pull/25999
• Own channels-last layout during backend construction by @glenn-jocher in https://github.com/ultralytics/ultralytics/pull/26001
• Restore automatic CPU channels-last selection by @glenn-jocher in https://github.com/ultralytics/ultralytics/pull/26002
• Speed up PIL and ndarray prediction sources with cv2 color conversion by @JESUSROYETH in https://github.com/ultralytics/ultralytics/pull/25988
• Fix plot_images scaling pose keypoint confidence with coordinates by @aleehydar in https://github.com/ultralytics/ultralytics/pull/24432
• Fix imread TIFF branch ignoring flags and extension case by @lszSamLin in https://github.com/ultralytics/ultralytics/pull/25986
• Fix CLI classes filter being dropped for world and yoloe models by @chinazhangchao in https://github.com/ultralytics/ultralytics/pull/25882
• Skip OpenCV's per-pixel grid in BoT-SORT motion compensation by @JESUSROYETH in https://github.com/ultralytics/ultralytics/pull/25995
• Add type checking for dataset YAML by @Y-T-G in https://github.com/ultralytics/ultralytics/pull/21905
• Fix Matplotlib error when restoring original backend by @Y-T-G in https://github.com/ultralytics/ultralytics/pull/21863
• Fix concurrent MongoDB tuner iteration IDs by @glenn-jocher in https://github.com/ultralytics/ultralytics/pull/26000
• Fix PyTorch 1.9 backend initialization by @glenn-jocher in https://github.com/ultralytics/ultralytics/pull/26003
• Improve Tuner search with confidence-weighted covariance by @glenn-jocher in https://github.com/ultralytics/ultralytics/pull/25996

New Contributors

• @chinazhangchao made their first contribution in https://github.com/ultralytics/ultralytics/pull/25882
• @aleehydar made their first contribution in https://github.com/ultralytics/ultralytics/pull/24432
• @lszSamLin made their first contribution in https://github.com/ultralytics/ultralytics/pull/25986

Full Changelog: https://github.com/ultralytics/ultralytics/compare/v8.4.135...v8.4.136