Correct method to prepare input data/shape for model training using keras?

Question

I have read through keras documentation and have checked countless blogs to solve this issue, but I remain puzzled.

I created a variable \'x\' that stores all my imag