Question on data augmentation

Hi, I read the chapter 2 and I have tested with my own dataset.When testing because I am interested in the shape of the image set, I should not deform the image, instead I have found some options like FlipItem or Rotate, to avoid the deformations provided by aug_transform. When applying them to batch_tfms I have the following doubts:

Am I augmenting the data or just transforming them?
If I am not augmenting them, what would be the correct way to use, for example Rotate, to augment the data, without deforming it?

Oscar,

Augmenting and Transforming in this context are the same.
You are taking a single image that is provided and altering it a little bit (i.e. transform) so that it looks like you have different images during each epoch.

See the docs here https://docs.fast.ai/vision.augment.html#aug_transforms to see all the parameters.

I think adding max_warp =0.0 , p_affine =0.0 would keep the image from defoming.