New: Ultralytics YOLO-Human

[Laughing-q]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Introducing new YOLOHuman model for human attribute detection! 🚀

📊 Key Changes

• Added YOLOHuman class as part of the model imports.
• Introduced a new YAML configuration for the YOLOv8 human detection model.
• Implemented additional augmentations to include human attribute data handling.
• Established a new dataset class, HumanDataset, for loading and processing human-related datasets.
• Included Human object in results to encapsulate detected human attributes.
• Enriched model __init__.py to include YOLOHuman.
• Formulated HumanPredictor, HumanTrainer, and HumanValidator under the new YOLO human module for prediction, training, and validation.

🎯 Purpose & Impact

• Enhances Model Catalog: Expands Ultralytics' model offerings to include human-specific attribute detection.
• Improves Dataset Handling: Offers streamlined process for datasets involving human features.
• Facilitates Human-centric Applications: Paves the way for more sophisticated applications such as demographic analysis, security enhancements, and personalized customer experiences.

[glenn-jocher]

@Laughing-q @Burhan-Q I attempted to de-conflict this with YOLOv10 in bba1152, I saw a mismatch here with m.reg_max and m.reg_max - 1 in v10, so went with the v10 version for the de-conflict, not sure if I might have broken something here in the training or if the Human models need retraining with the this new argument changed:

[Laughing-q]

@glenn-jocher No worries! I've updated it back to reg_max.
BTW the reason why I made the update reg_max-1 to reg_max here is that I found we have everywhere reg_max equals to 16 expect the one in BboxLoss(because we are passing reg_max-1 to initialize the loss function). I figured we should keep the consistency everywhere for reg_max for better maintenance.