AutoBatch report include reserved+allocated

[glenn-jocher]

May resolve #9287 (comment)

Signed-off-by: Glenn Jocher [email protected]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Improved the batch size estimation in YOLOv5 for better memory utilization.

📊 Key Changes

• Clarified the comment to specify that it's for YOLOv5 batch size estimation.
• Updated the formula calculating the fraction of CUDA memory used to include additional terms.

🎯 Purpose & Impact

• To allow users to better understand that the autobatch function is tailored for YOLOv5.
• To enhance the accuracy of estimated batch size, thereby improving memory usage efficiency which could potentially reduce the risk of out-of-memory errors and increase performance.
• Users could see improved results in terms of training speed and system stability when training YOLOv5 models.