Both only cover the math as it becomes relevant. I am working through fast.ai's book right now and find its pragmatic approach to DL pretty agreeable to just getting models hosted and out the door. I watched the lectures before hand, and there are several Jupyter notebooks and examples on how to get models deployed ASAP with clunky interfaces, which also might be of interest to you.
Both only cover the math as it becomes relevant. I am working through fast.ai's book right now and find its pragmatic approach to DL pretty agreeable to just getting models hosted and out the door. I watched the lectures before hand, and there are several Jupyter notebooks and examples on how to get models deployed ASAP with clunky interfaces, which also might be of interest to you.