Data Augmentations?

[Elizabetht1]

Hi! I wish to apply random augmentations to a small dataset. I have two strategies in mind.

1. Each time a sample is drawn, apply a different set of random augmentations. This way, even if a dataset only had N samples, if it was trained for 50 epochs, it would see 50*N different images
2. Apply random augmentations to the dataset before training, to increase its size. So, if there were N images and 50 different augmentations to apply per image, you would have 50*N images going into the augmentation process.

Is one of the above better supported in monai label? If not, what is the best way to perform augmentations in a way that increases the size of the dataset?

Thanks!