Faster threaded plots

[glenn-jocher]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Refactor threading and minor plotting fixes in YOLOv5's utility modules.

📊 Key Changes

• Moved the threaded decorator function from general.py to __init__.py to centralize its definition.
• Added threaded decorator to several plotting functions within metrics.py and plots.py to allow them to run in separate threads.
• Changed plt.title to fig.title and plt.close to fig.close in plotting functions to properly reference figures and close them.

🎯 Purpose & Impact

• By centralizing the threaded definition, the codebase becomes cleaner and reduces redundancy.
• Multithreading the plotting functions can improve the overall efficiency by allowing plots to be generated without blocking the main execution thread.
• Understanding and maintaining plots become easier with clearer references to figures.
• Users may experience faster plotting and analysis when working with YOLOv5's visual tools, especially when dealing with large datasets or complex visualizations. 🚀

[glenn-jocher]

Superseded by #9154