The Course was more about the the types of neural networks and how they work than Tensorflow, except for week 1 where we had a Tensorflow introduction, I could gather a lot from the programming assignments but I think there needs to be more about the Tensorflow library in the lectures.

I found the practice notebooks of this course to be lacking due to two reasons: 1) The notebook links are broken, resulting in my not being able to complete them. 2) The notebooks do not have practice sections where we could code ourselves following the examples given.

The course is good, but 1) the lab environment is not working at all.... I had to run the notebooks on google colab ! 2) The code is outdated. Tensorflow 2.x is out.

Overall good course but lectures were a bit weak on underlying math, compared to labs which made it a challenging at times to tie the two parts together.

One of the better courses in the IBM AI certificate. The notebooks are nicely annotated and have more relevant information than the video lectures.

All the code is TensorFlow1, which is unfortunately completely outdated. Also no assignments or final. But good for the fundamentals of TF.

Basic level but well explained, useful notebooks, not much on Tensorflow, more on the theory of the networks. Uses outdated Tensorflow v1

The course is good but you have to change the codes from TF1 to TF2 since is dificult for the learner tranaslate de codes by himself

I do not consider the practical part to be exactly beginner level, but the theoretical material is very good.

Labs would need to be updated to new versions of Tensorflow

The presentations were clear and concise

There is a sense to be outdated. Not much activity in the forum, code which is not updated...

It would be better if the exams are a bit more tough.

The questions are too easy to solve.

Too focus in coding but not in the underlying concepts and how to use the libraries.

Interesting view on tensor flow, but gap between labs and videos is quite big.

Low explanation.

But in this short duration we can't expect more.

Course needs to be updated to Tensorflow 2.0 at least.

we should have more videos and theory also..

Material excellent, cramed though.

The tensorflow version is outdated

Outdated and not in depth enough.

Falto mas teoria

No real dataset

no applications

Disappointing stuff. The videos teach complex stuff like recurrent neural networks (RNNs like LTSM), restricted Boltzmann machines, and autoencoders very quickly - less than 10 minutes per "week" of learning. While the labs are extensive, you don't learn anything as the amount of TensorFlow code is totally intimidating and none of the steps are really explained. You can copy the code, but you won't develop an understanding of it in this course. Not to mention the code is so heavy the Skills Lab times out before the network is trained. Still, if you just want to claim you've done Tensorflow, you can click through the stuff in about 30 minutes per "week" of learning.

The lectures only cover various neural nets and not how to actually implement them on Tensorflow, which should be the gist of the course. Further, the labs are at many places not compatible with the most recent Tensorflow version 2's, and only work for previous Tensorflow version 1's which are quite different. The labs must be re-written for the newest versions given Tensorflow's backward incompatibility.