train.py --logdir argparser addition

[glenn-jocher]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Improvements to logging and path handling in train.py script.

📊 Key Changes

• 🛠 Refactored directory path handling using Path from pathlib for better clarity and consistency.
• 📁 Changed the log_dir assignment to support a custom logging directory passed via the new --logdir argument.
• 🚀 Ensured creation of the logging directory (log_dir) and its subdirectories.
• 📝 Updated the path concatenation method to leverage the / operator provided by pathlib for joining paths.
• ✨ Added a new --logdir command-line argument to specify the logging directory.

🎯 Purpose & Impact

• ✅ Ensures clear and flexible specification of logging directories, improving customizability for different runtime environments.
• 🧩 The use of pathlib provides a more readable and robust way to handle file paths, which can prevent path-related errors.
• 📊 Users can now choose to save logs in a specific directory, enhancing the organization of training runs and making it easier to manage experiment artifacts.
• ➕ Enhancements may lead to a smoother user experience and greater ease of use for both new and existing users of the YOLOv5 model training process.

[glenn-jocher]

Addresses issue #640