Does it matter if you take Course 5 before Course 4?
No, if you are already familiar with TensorFlow and don’t care about learning CNN.
2 Likes
Yes, I agree with Saif. The architectures of CNNs and RNNs are independent of each other, so there’s not really any material in DLS C4 that is a prerequisite for C5. But C4 is the first course where we really get serious about using TensorFlow and C5 assumes you are already familiar with how to use TF in a reasonably sophisticated way. If you have other experience with TF beyond the very brief intro that we got in W3 of DLS C2, then you can do C4 and C5 in either order.
1 Like
Thanks. That would be my next question… we just touched on TF earlier in a C2 assignment. Are TF methods applicable to some of the other basic Python coding we have already been through? … for example, implementing CNN or RNN on your home computer?
Sure, algorithms implemented in TF are portable and you can run TF on your own computer.
Here’s a thread with instructions for setting up at least one kind of local environment, although a lot of the complexity there is in trying to actually duplicate the versions of the online environment. When you’re doing your own thing, you don’t have to be so concerned about the version duplication part of things, although it’s always good to have something like Conda in your toolkit.
Note that if your only exposure to TF is Week 3 of DLS C2, then you are skating on pretty thin ice here to jump straight to C5. It depends on your level of experience and your learning approach. If you are pretty experienced at debugging in general and using complicated software packages that use OOP, then you’ll be fine if you take advantage of the documentation on the TensorFlow website or looking at StackExchange when you hit a problem. There are lots of tutorials on the TF website. But if you’re relatively new to all this and found the material in DLS C2 W3 at all challenging, then it might be wiser to just go ahead and take C4 first. C4 has plenty of really great and important material. Of course CNNs are important for lots of image and vision related problems.
Or you can just dive into C5 and see how it goes. You’ll know pretty quickly whether it’s working for you or not. 