I took the Deep Learning Specialization before (which admittedly set a very high standard) and expected this specialization to be similar, just more focused on Tensorflow. I couldn't have been more disappointed. Explanations are very shallow, and I totally missed the well-thought-out programming exercises. Here there are just notebooks with some missing parts that you have to fill in, without any explanation. You just copy over the code from the lectures, and it works, and you have not learned anything. A sneak peek at the next course shows that this seems to be organized in the same way. Luckily I finished this course within the 7-days trial...

I can't believe I paid $49 for this and knocked it out in <2 days. I now know how to copy-paste lines of tensorflow to do some very specific things. It's fine to not go in-depth on the math and everything behind the scenes if you instead focus on practical application, but this course does neither.

For a glorified tensorflow/keras tutorial, I would hope that after the course I would be able to build my own models, but I honestly know how to do nothing outside of the 6 lines of code I copied from the videos, and I don't even know how those lines work.

Disappointing course. It is an introduction to high-level Keras API and better be called this way. For anything else it is not useful. An introduction to AI, ML - really? :) It does not explain how do neural networks actually work and what is going on when fit() is called, etc. How about about gradients, optimizers, activations, etc? Only convolutional networks are properly explained. Labs are copy-paste from slides and zero thinking. I would better not comment on the quizzes.

The course does not goes deep (not even close) into explanation, and many topics/methods are just mention. Per se, that is not bad if you could use the forum to get the answers of the questions that arises. However, that is not the case for this course, where most of the questions are unanswered.

Moreover, there are many bugs/typos/mistakes in the given code. And because the staff does not answer the forum's questions, you are not sure if they are really bugs or not.

Very very very superficial course, both on the theoretical part and on the practical part.

The theoretical part is almost none. Each week you will have like 2 minute video explaining the NN that you will use.

The practical part is also very superficial because the only thing that you do is calling functions from Keras.

Don't recommend this course to people that already have some (very little) experience with tensorflow/keras. Recomend it to newcommers

Jupyter notebooks and video explanations have many issues (for example: https://github.com/lmoroney/dlaicourse/blob/master/Course%201%20-%20Part%204%20-%20Lesson%202%20-%20Notebook.ipynb - where fundamental terms like "accuracy" and "loss" are mixed up). Considering that many people who will be taking this course may be beginners, such mix-up can really affect what they learn and how they approach Deep Learning.

Many in the course were just taught superficially. No in depth clarification of things was given according to me. Rather I would recommend to check out Imperial College London's course "Getting started with Tensorflow 2" course which is awesome. There are 2 courses in the specialization, I have mentioned first one of them check out other(s) on your own. Happy learning.

.

Course is interesting.

But programming assignments are not described enough. Leaving students with difficulties to understand exactly what is expected to get the grades. Even with a working solution, multiple assignments must be send to try to "fine-tune" the programming assignments to get a grading without knowing why it does not work.

It should not be called like it is called. This is not even close being "an Artificial Intelligence, Machine Learning, and Deep Learning". At most this is like "10 minutes into TensorFlow". You have to be at the absolute beginning to get out something from this course. Better read the docs and take A. Karpatny course on youtube

Honestly, this course doesn't meet my expectations. The instructor's way of explaining the code was okay .This was really a very easy course. The topics covered in this course can be studied from any youtube channel with free of cost. If you are wishing to get a deeper understanding of tensorflow then don't buy this course.

The lections were good enough, but the assignments are awful. In the first week, they say we can use TF 2.0, the Google Colab uses TF 2.3, the grading script accept only TF 1.14. I spent a lot of time only to figure out which version syntaxis should be used in order to pass the grader. That was awful.

Not going to lie, I'm pretty disappointed after finishing Andrew Ng's course and coming here for the next step. I feel like maybe Andrew raised my standards too high, but this course was extremely not begineer-friendly as an intro course, and I don't think I was able to pick up anything worthwhile.

I have been a huge fan of the DeepLearning.ai specializations and content. This specialization is rough around the edges and really only scratches the surface. If this was the only content I had seen from them I would be left wondering what the heck I am learning and have no idea how it works.

My feeling is that this is not a whole course but just one or two weeks that have been taken from a course. The contents are really basic, the quizzes are mostly trivial, and the exercises are mainly about rewriting fragments of code that they have already given to you in the labs.

I had previously done the deep learning course of deeplearning.ai. I expected this specialization to build up on it, but it rather seems to be a step back.

For me the course was to good to rush through it and just pick one or two new ideas. Hope next courses are better.

The professional content is very poor - it cover very little and only the very basic meaning is taught . I have expected a much more in depth technical learning and a little more theory. All the course taught is how to write 20-30 lines of code... disappointing.

Good course content 5/5 for it. However, the last exercise is prone to errors due to the system. There are many posts in the forum and none of them have a response by the staff. Considering this is paid content I'd expect staff to address the problems.

The code in googlecolab was not organized well and the course content was so basic. I expected more from this course when I started as I took course of ML with Andrew Ng and expected the same level of quality but got disappointed.

Amazing lecturers and good structure thus 2* but the content would be 1* because it is far too basic and this certificate is unusable to show anywhere because of that. The tasks one up to 3 hours to accomplish it? 5 min max...

Introduction to Keras would have been a more appropriate title. A bit disappointed, that the course just scratched the surface of tf.keras instead of explaining more concepts of the (much bigger) tf library.

I am having issues resetting my course deadline... It keeps saying this - Convolutional Neural Networks in TensorFlow

0No upcoming sessions are available. Please check back later.

what do I do?.

the assignment submission is rediculous. in 37 tries surely my code is right somewhere along the line.

especially seeing as i get the right result every single time.

assignment 1 has bugs

No theory at all. Not much explanation regarding TF classes, functions and their arguments. Just basics of TF. Most of the assignments can be solved with copy/paste from examples.

Just a brief introduction to TensorFlow, very basic and short on practical exercises. I literally copy-pasted texts from one notebook to another. Neither gives it a lot of theory.

Programming exercises are quite sub-standard. Explanations in video lectures are too short and coarse. Andrew Ng's deep learning specialization far superior, stick with that