Hello! Thanks for contributing to the discussion with such thoughtful suggestions. 🌟

Displaying the number of images per category during evaluation sounds like a reasonable enhancement for gaining more granular insights into model performance across different classes. This can definitely help users better understand the specific areas where the model excels or requires more training data.

If you are willing to submit a PR for this feature, it would be a fantastic addition to the project. Please ensure your changes maintain the current functionality as an option, perhaps through a user-defined flag in the evaluation script. This way, users can choose the output format that best suits their needs.

Looking forward to seeing your contribution!

@glenn-jocher I have implemented this feature and used the per_class_img parameter to select whether to output the number of images for each class. I would appreciate some suggestions for better parameter naming.

Hello @sunmooncode! Great to hear that you've implemented the feature. 🚀 For the parameter name, how about detail_per_class? It's concise and clearly indicates that additional details will be provided for each class. Let us know if you need further assistance!

@glenn-jocher Very well, I will update it!

@sunmooncode thanks for the update! Looking forward to your PR. If you need any help, just let us know! 😊