Refactor for simplification 2

[glenn-jocher]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Improvements to file handling and logger callbacks in Ultralytics YOLOv5.

📊 Key Changes

• Simplified reading of JSON files in export.py using Path.read_text().
• Streamlined the callback conditions for model saving related to Weights & Biases (wandb) and ClearML in utils/loggers/__init__.py.
• Ensured the best model is saved at the end of training with a check for evolution in ClearML logger.
• Updated the on_params_update method with type-hinting for better code clarity.

🎯 Purpose & Impact

• Code Simplification: Using Path.read_text() enhances the clarity of the code by removing an extra file opening step.
• Logger Refinement: The changes reduce code repetition, making it tidier and possibly easier to maintain.
• Best Model Preservation: Adjusting the saving of the best model ensures it is not overlooked during training completion, particularly when not using hyperparameter evolution.
• Type Hinting: Adding a type hint increases maintainability and helps developers understand the expected input types.

These changes should make the codebase more robust and user-friendly for developers integrating YOLOv5 into their projects. 💻🔧