Data Augmentation

I am using detectnet 2 and using some of the augmentation provided in the config file, I am curious as to how is this translated into training.

Does the dataset augmentation 2x the data or only change the current dataset provided?

If spatial augmentation is enabled, spatial transformation matrix (preprocessing + augmentation) will be computed.

So it does not add any images to the dataset, but only augments the input data?

Correct.

Thanks