Error: google.protobuf.text_format.ParseError: 57:5 : Message type "AugmentationConfig" has no field named "transform_prob"

Since the blog is released half a year ago, it is for 3.0-dp-py3 release.
Currently, the latest docker is 3.0-py3. There is some change in augmentation part.
So, please follow NVIDIA TAO Documentation to modify the augmentation part of your training spec file.