Daemon plot_labels() for faster start

[glenn-jocher]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Enhanced threading in callbacks and adjusted label plotting in training.

📊 Key Changes

• Removed plot_labels call from train.py to streamline the training process.
• Updated callbacks to potentially execute in daemon threads for asynchronous operation.
• Simplified download_one comment in utils/general.py.
• Integrated label plotting directly into a callback for cleaner management.
• Minor adjustments to comments and code organization for clarity.

🎯 Purpose & Impact

• 👍 The removal of immediate label plotting during training reduces initial overhead and allows for a more focused start to the training process.
• 🚀 Allowing callbacks to run in separate threads can improve performance by not blocking the main thread, making training more efficient.
• 📈 The integration of label plotting into a callback could make it easier to manage custom logging or plotting routines.
• 🧹 These changes lead to cleaner code structure and potentially better resource utilization during training, positively impacting users' experiences with YOLOv5 training sessions.