rename class autoShape -> AutoShape

[developer0hye]

follow other classes' naming convention

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Improved naming convention in YOLOv5 codebase.

📊 Key Changes

• Renamed autoShape class to AutoShape for consistency in class naming.
• Updated constructor and relevant print statements from autoShape to AutoShape.

🎯 Purpose & Impact

• 💅 Code Readability: Harmonizes class naming conventions for better clarity, adhering to Python's CapWords convention.
• 🛠️ Refactoring: Reflects best practices in software development, but no changes in functionality.
• 🔍 User Transparency: Clearer logging messages to inform users when AutoShape is enabled.

[glenn-jocher]

@developer0hye thanks for the PR! This is not a bad idea to stay consistent with the pytorch camelcase module convention. It looks like there are 6 case-sensitive occurences of 'autoShape' in the repo, but the PR only includes 4 changes. Can you do the other two also please? Thanks!

[developer0hye]

@glenn-jocher I modified them! Please check my commit~

[glenn-jocher]

@developer0hye looks good! Thank you for your contributions!

[developer0hye]

@glenn-jocher

Hello, jocher.

I measured the execution time of 300 over imagenet pretrained backbone networks.

Refer to https://github.com/developer0hye/pytorch-backbone-benchmark

I hope it helps you to find more efficient backbone than current yolov5's backbone model!

[glenn-jocher]

@developer0hye wow, that's impressive!

You should definitely plot your results a few different ways, maybe using plotly or an interactive dashboard so you can mouse over the best datapoints and get their model names etc.

top1 vs time
top1 vs params
params vs time

[developer0hye]

@glenn-jocher

https://colab.research.google.com/drive/1NmkUIcA9Vt8U2WLXgZeIrIuY3yFW3Byx#scrollTo=9NyYQqA_iwzn

Look at this!

[glenn-jocher]

@developer0hye very nice!