Data augmentation methods used in pre-trained models provided by torchvision for image classification?

Hi, I’m interested in using the pre-trained models provided by torchvision for image classification. I was wondering if you could tell me what data augmentation methods are used in these models? Specifically, do they include any color-based random augmentations? Thank you!

All available weights from pretrained torchvision.models are available here including their training recipe. You could check the commands used to train each model and check how the train.py script is called using these arguments.

Thank you for providing an answer to my question on the post. Your response has been very helpful to me.