This course significantly lacks depth. The topic is covered at a very high-level and represents only a lightweight introduction. You will not gain any insights into the challenges that someone might face using CNNs on Tensorflow in a real-world scenario.

This course does not compare to the kind of insights that you learn from the other courses taught by Andrew Ng.

There are no graded programming assignments to validate what you have learned. The exercises that are provided are very simplistic.

You may look at it as a set of use-cases on how to work with particular types of .ipynb notebooks or how to structure your code, but, unfortunately, lectures are useless and tasks are mechanical rather than challenging.

Huge disappointment.

I think I knew more about CNN before this course.

This is the second course of the specialization and still I feel like I haven't been introduced to anything beyond the free tutorials available on tensorflow website. So far the specialization has also been only focused on the keras api of tensorflow which makes me feel that perhaps the name of this specialization has been poorly chosen (perhaps it should be 'Keras in Practice Specialization'). On the positive side, the instructor is eloquent and the learning material is presented in a well and orderly fashion (ignoring some minor cases of redundancy in notebooks; basically copy pasting the whole notebook several times just to introduce a few lines of new code).

This is pretty simple. This doesn't give an idea of the real use of keras. also there is no programming assignments.

This course is extremely disappointing. The content is very shallow, you'll get more from just following the keras tutorials in the official tensorflow docs. Also, since this specialisation only seems to cover the keras api, perhaps the title is a bit misleading.

On the plus side - it is pretty easy to complete the whole thing in a day and very easy to knock it out before the free trial ends. But honestly, even for free, I don't think it is worthwhile.

The material is very presented in small repetitive chunks, where you'l basically just be running the same notebooks over and over with one small new function thrown each each "week". The quizzes and assignments are riddled with typos which I think is a poor show for a paid for course.

The assignments are basically just copies of the coursework notebooks. No thinking required.

I really would not recommend this specialisation. Your time will be better spent elsewhere. It is such a pity as the previous courses by Andrew Ng have been of such high quality.

This course is so short in content that, in the whole last week, it is explained a trivial concept about multi-class classification. Besides, the last quiz recycle questions from the previous quizzes from this and the previous course. It is clear that the course was made in a hurry once the notebook examples lack in written content or figures explaining the subject. Finally, there is no practical assignments in this "Tensorflow in practice" course.

This series of courses is just a 'Hello World' introduction of Tensorflow/Keras. The instructor just touches the surface of some code from the Tensorflow document without explaining some really fundamental concepts (e.g. tensors). The videos are usually 1-2 min long, really a headache to watch. The quiz is too simple and poorly designed-- instead of thinking or calculating you just need to remember some basic concepts/grammar rules. Programming exercises are not really useful and there is too much duplicate work. Not worth the money if you plan to pay for it-- auditing is enough. I am disappointed by deeplearning.ai for producing such a shallow course.

I finally feel confident that I understand the basics of Convolutional neural nets and what function the various layers serve. It took a Polymath computer engineer/science fiction writer to finally break that mental block and get through to me. Take this class you won't regret it.

A very comprehensive and easy to learn course on Tensor Flow. I am really impressed by the Instructor ability to teach difficult concept with ease. I will look forward another course of this series.

this course is very useful for beginners

It is a great intro but a very limited course. Short videos and a small number of examples, for example, Transfer learning could be more in-depth. Week 4 really made a few obvious changes in the code. I do think it's great material, but all of it could be made into a 2-week course instead. Thanks for your efforts.

An excellent course by Laurence Moroney on explaining how ConvNets are prepared using Tensorflow. A really good strategy to have the programming exercises on Google Colab to speed up the processing.

I enjoyed doing this course on CNN in Tensorflow. Thanks for the lectures by Laurence Moroney. And it is always a pleasure to hear Andrew Ng explain even difficult concepts in simple terms. He is one of my favorite teachers online, and reading about his ML course in a New York Times article back in 2012 or 2013 made me completely change my career direction and motivated me to eventually get into cloud and Big Data! And thanks also for the exercises on codelab. That makes it really convenient to learn and experiment with Machine Learning and Deep Learning.

I did take the first course in the Deep Learning Specialization early last year, but didn't get a chance to do this until now. Looking forward to completing the remaining three courses sometime this year.

This course follows up with two very important concepts that are left out in the first course of the series.

While the workload is definitely not heavy, the quality is high and the explanations are top-notch.

This whole specialization focuses on practice, it helps you understand little by little the building blocks to create a model. If you want a theoretical explanation sign up for the Deep Learning Specialization (as suggested by the instructor itself).

This specialization should be seen as complementary to the other, expecting to find the same concepts explained over and over again wouldn't make any sense and it would only be redundant.

It's an amazing course, the video lectures are fruitful and the contents of the courses are well designed, the instructor is talented and his explanations are extremely helpful, it's one of the best courses taught on Tensorflow!

Last assignment could have been explained better. Laurence does not talk about ImageDataProcessing.flow, despite this is the only solution

Throughout this course, I quickly fell in love with the new concepts introduced by the tutors, especially transfer learning which allowed me to skip the hard part of the training process by acquiring previously trained models on larger datasets, thus guaranteeing more accurate predictions and better results. In addition to that, I liked the idea behind dropout regularization, one of the strongest techniques to reduce interdependence between neurons and consequently minimize overfitting. In the end, I got impressed by the shift from binary to multi-class classification and the fact that we no longer have to limit our results to two different output classes but to many more. I am delighted to be a part of this community and I hope to extend my learning journey on Coursera.

The course initially revisits basic concepts of convolutions, model compilation and building. The concepts that are built on this include image augmentation, transfer learning with Inception net and multi class classification proficiently cover practical implementaiton using TensorFlow