How confusing would it be to skip the CNN course, and just go into the Sequence Models and then go back to the CNN course afterwards? Could I still understand what’s going on in the Sequence Models course?
I do know a tiny amount about CNN like pooling and some filtering convolutions, but not really in depth.
I have not done the whole Sequence Model course, only the first two weeks, but my take is that there isn’t really any cross over in terms of the technical material about ConvNets versus RNNs. So in that sense, you can do the courses in either order.
But there is another issue: if you jump straight from DLS C2 to DLS C5, you don’t really know very much about TensorFlow, unless you already have your own separate experience with TF. In DLS C4, we get a lot more experience with the more sophisticated features of TF as we proceed through.
So if you already know TF, then it would be fine to take C5 and then C4. If this is your first encounter with TF, then I think you’re asking for trouble, but there are probably ways to cope. If you are aware of the risk, then spend some time on the TF website reading tutorials and look at useful threads on the Forums, e.g. this one which explains the Sequential and Functional Keras APIs.
thank you!