Fix float zeros format

[glenn-jocher]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Updated number literals in division to integers for consistency across codebase.

📊 Key Changes

• Unified the division operation by using integers (255) instead of floats (255.0) where values are scaled.
• In some instances, changed from a floating point number with a decimal (1.) to an integer (1) when used in division or as a tensor value.

🎯 Purpose & Impact

• Consistency: Ensures consistency in using number literals throughout the codebase, avoiding unnecessary type casting.
• Clarity: Using integers for division by 255 when normalizing image pixel values provides clear intent and understanding for developers and users.
• Performance: May negligibly improve performance by using integer division where applicable, though modern compilers may already optimize such cases.
• No Functional Change: Should not affect the output or the overall functionality of the models, thus the impact on end-users should be minimal or none.