Update autobatch.py

[glenn-jocher]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Enhanced memory reporting in YOLOv5 autobatching with added device name and more precise metrics.

📊 Key Changes

• Now includes the device properties, specifically the GPU name, in the memory reporting output.
• Memory metrics (total, reserved, allocated, and free) are now reported in GiB (Gibibytes) with two decimal places for precision.
• Formatted the print statement to improve readability and information detail.

🎯 Purpose & Impact

• This update is intended to give users a clearer understanding of their system's GPU memory usage during the autobatching process.
• Improved precision in memory reporting can help users make informed decisions when allocating resources for model training and inference.
• Including device name will immediately clarify which device's memory is being reported, useful for systems with multiple GPUs.
• Overall, these enhancements can contribute to better resource management and potentially smoother and more efficient model operations.