Refactor modules

[glenn-jocher]

Follows 'Refactor collections and fstrings' #7821 for python files in models/

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Refactoring and code cleanup in Ultralytics YOLOv5 models.

📊 Key Changes

• Simplified the loop in the forward function of a neural network module using list comprehension.
• Streamlined condition check for Detect layer initialization to improve code clarity.
• Removed unnecessary else-clauses after return statements for cleaner code structure.
• Enhanced TensorFlow model's layer enumeration by eliminating redundant iteration indices.
• Optimized TensorFlow's predict function by ensuring that only relevant return values are used.
• Improved variable naming and data processing in dataset generation for model quantization.

🎯 Purpose & Impact

• 👍 Readability: The changes made to list comprehensions and condition checks make the code easier to read and understand.
• 🚀 Code Efficiency: The removal of unnecessary variables and indices can lead to slight improvements in code execution efficiency.
• 💡 Clarity in Outputs: Changes in TensorFlow prediction outputs ensure clarity on what is being returned, which benefits developers debugging the model.
• 🧹 Cleaner Codebase: A cleaner codebase eases maintenance, further development, and lowers the barrier for new contributors to understand the code.
• 🛠 Better Tooling: By refining the quantization dataset generation, the models are better equipped for conversion to optimized formats compatible with various hardware for improved performance.