Update hubconf.py to reset LOGGER.level after load

[glenn-jocher]

Resolves silent outputs after model load.

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Improve logging behavior during model creation in YOLOv5.

📊 Key Changes

• Captured the initial logging level before changing it.
• Restored the original logging level after model creation process.

🎯 Purpose & Impact

• Purpose: Ensures that the user's logging level preferences are respected during model loading, making the library more considerate of the user's logging settings.
• Impact: Users will experience less unwanted logging output when using YOLOv5 in their projects, leading to cleaner and more readable logs.

🔍 This update is particularly useful for developers looking to maintain a specific log verbosity within their applications or for those integrating YOLOv5 into larger software systems where log output needs to be controlled.