[FIXED] Dataframe reshaping for tensorflow training

Issue

I'm trying to get some data ready for training in a GAN model. This data is around 37k faces of resolution 28x28 with 3 channels each. At the moment, this data is in a Dataframe object and I would expect the shape of this data to be (37k, 28, 28, 3), but the shape turns out to be (37k,). Is there a way I can reshape this data to conform to this data to the proper shape?

Solution

What you want to do can be implemented using the stack function.

Answered By - Ben Grossmann