Tips on data augmentation in mxnet [R package]


From online examples and available documentation, it is rather straightforward to use mxnet for image recognition tasks in R. Ingenious work!

Most examples in R however skip the crucial part of data streaming & augmentation of image files (e.g. using mx.img.ImageIter). Are there any tutorials on doing this fully in R? I am especially interested in regression settings, i.e. where the target is continuous.

Any help is appreciated very much, even if the answer is “you need to do it in Python”.