While the bottom up approach to ML has its charm. Its hard to get to great results by trial an error that way.
So after Ng I suggest the fast.ai deep learning course which uses a top down approach. Much more practical if you actually want to get something meaningful done.
While the bottom up approach to ML has its charm. Its hard to get to great results by trial an error that way.
So after Ng I suggest the fast.ai deep learning course which uses a top down approach. Much more practical if you actually want to get something meaningful done.
I agree.