Image processing and Compilation info

[ansariahmad]

I trained a model for sign language. Then I write down whole code of model in python. Now I want to know that how image preprocessing is being done on images and what are the arguments of compilation and training functions of classification model. But it shows that compilation info isn't available. Is there any way to get all of this info???

[michaelrcrowl]

When you say "image preprocessing", are you asking about data augmentation (like random rotations or flips)? I came here to see if anyone else had a similar question, because I'm also quite curious given some things I'm trying with Teachable Machine. The underlying MobileNet doesn't appear to preprocess data (like in a Sequential layer), and the source code here always has flipped=False flags in some of the functions. As far as I can tell, Teachable Machine is NOT augmenting the dataset, but I'm really not sure.

[michaelrcrowl]

#3
I'm not sure if this older thread helps answer your question.