A few questions regarding specific Specializations (TF, NLP, GANs)

So, I have just finished the DLS and have been thinking about where to go from here. Also, I realize if I receive replies from Mentors I’m probably not going to hear anything ‘negative’… But I am just kind of seeking an honest assessment.

One thing I definitely feel I need is to get a better handle on TensorFlow. Though it was introduced in the DLS, I can’t say I yet I feel fully comfortable yet using it outside of the scope in which it was presented.

To that end, on the one hand I recently purchased a second hand copy of Francois Chollet’s Deep Learning with Python, 2nd ed. Written by one of the authors of Keras, one would expect this to be a pretty definitive resource.

That said, there are also the TF specializations here – A whole three of them in fact.

However, first of all, it is not clear to me if these three have any order around them, or aside from the Google Certificate course, the others are just more ‘topic oriented’ ?

At first, I thought the Certificate course would definitely be the one to have… But then I looked it up yesterday, and as of earlier this year, Google has suspended the TensorFlow certificate program for the time being. Which perhaps makes that course at least a ‘little’ less worth having…

Second, with regards to the NLP Specialization, while one of course would imagine it fleshes things out in a lot more detail, as it is an entire specialization, a good part of it seems to basically go over the same topics as DLS course 5-- In particular a lot of the material seems to be spent on RNNs and LSTM, which… These days does not seem as useful(?).

In any case, just to be honest, from a personal point of view… I guess I find NLP a little less interesting(?)-- Though, at the same time from an employment perspective NLP is all the rage these days.

I didn’t know if anyone had any opinions on that.

As to GANs, I think this would be interesting and would also be a good excuse to pick up Pytorch, seeing as employment is my primary goal at this point, I guess where might one rank this one in terms of priority given the other specializations above ?

Any advice would be greatly appreciated.


1 Like

Hmm… no takers on this ?

It seemed you didn’t want answers from mentors.

@TMosh oh no, anyone really-- I’d just like the answer to be honest, and Mentors can even DM me if they prefer.

I mean, for example, before I looked at the syllabus for NLP, I kind of presumed it was maybe 80% Transformers and their variants, and 20% everything else. I mean Transformers were brought up in DLS C5W4— but incredibly briefly… I mean I still have so many questions.

But without knowing, the way the NLP specialization is structured, it seems less a ‘continuation’ and more like-- ‘I don’t need to know about anything that came before and just want to study NLP for its own sake’.

I could be wrong.

Same with TF-- is there some ‘order’ here of the specializations ?

In any case, as I am more at the ‘electives’ point and it is my time and money make sure it is well spent when it could otherwise be diverted to different areas (i.e. not to take a class for the sake of ‘taking a class’).

I also have still been thinking about doing that Stanford Probabilistic Graphical Models Specialization (https://www.coursera.org/specializations/probabilistic-graphical-models).

Unfortunately, at the moment I feel I only have the time to choose one.