Question for mentors: Can Course 5 be taken before course 4 in the DeepLearning Specialization?
I am wondering if the material in course 5 builds on course 4. I need info recurrent NNs for a project that is going on now, so I am thinking about switching the order. Does it matter if one studies sequential NNs before Convolutional NNs?
Thanks!
In terms of the core material, the two courses are independent: RNNs and CNNs are really different and not related. But there is another level of consideration here: it is in Course 4 that we are introduced to a lot of TF 2.x and Keras concepts. Also the general complexity of dealing with CNNs is a lot higher than the networks we dealt with in Course 1 and 2, so it’s a good mental preparation for dealing with RNNs which are arguably even more complicated.
So I think the primary consideration is how much prior experience you have with TensorFlow and Keras. If you’ve used those before, then you’re probably fine to go straight to Course 5. If those are new to you, then you’re probably in for some rough sailing ahead. But feel free to give it a try and you’ll see pretty quickly whether TF will be a roadblock for you. As long as you’re willing to do a little googling and spend some time on the TF docs and StackExchange when you hit an issue, you’ll probably be fine to go straight to Sequence Models.
@paulinpaloalto, thank you for the valuable information!
I am new to TF so will stay with the recommended course order.
If you are literally seeing TF for the first time, the first introduction to TF in these courses is in Week 3 of Course 2. If you’re already paying the big bucks to sign up for everything, it might be a good idea to at least view the lectures that introduce TF in Week 3 of Course 2 and maybe take a look at the assignment. I don’t think you need the other previous material in Course 2 in order to understand the TF assignment. It’s more or less independent, but you might want to watch the lectures about softmax as well since that’s what the assignment uses.