First, thank you for the great project!
I'm planning to fine-tune the model on my custom dataset and would appreciate some guidance on the amount of data needed for reasonable performance.
I already saw you use 500 images per class to train MVTec dataset.
Could you offer a rule of thumb for the minimum number of images per class you'd recommend?
For example, are we talking about 100, 500, or 1000+ images per class for a task like this?
Any advice would be a great help.
Thanks!