As I have gone along in the deep learning specialization courses I am noticing more and more places where the video was not edited cleanly. For example, Andrew will repeat a sentence multiple times assuming his error will be edited out later, however there are many examples of these edits not happening

Fantastic, but it would be more reasonable if the courses from this specialization was more "hands-on". I feel like I can implement nothing by myself after this course without looking and stealing the code simultaneously. But other than that, I would love to recommend this course to any of my friends.

This is yet another splendid course on CNN by Andrew Ng, simplyfing the concepts of Convolutional Neural Networks to a newbie. The reason I'm not giving it a 5 star is because the I got stuck in a couple of instances in assignments and when I looked at the resolution, I felt it could've been avoided.

The explanation is more than great. My only note is regarding the assignments. I thing there should be a preparatory course to be mentioned at the beginning of this course to learn tensorflow. I am not comfortable doing my exercises by following the functions that I need to call from tensorflow.

Keras should have been thought, and few functions were tough to crack , for example [:,:,:,c] for the cth filter . such stuff have taken a lot of time. Theoritically it was damn good. But faced little tough with programming assignments, but was good to brush up with challenging stuff like these.

The information and exercises contained in this course are great. However, there are a lot of video editing bugs that distracted me while I was watching. Many places have repeated phrases, where the instructor obviously meant to correct a previous statement, that were not properly edited out.

The course is an excellent introduction to convolutional neural networks. As always, in limited time, there is a trade off between depth and breadth. What content and assignments there were were excellent. Probably the only reason why I rated the course with 4/5 stars is that I wanted more.

This course touch many state of the art deep learning networks especially in image processing. The programming assignment is more challenging to me mainly because I am not that familiar with tensor flow, and higher number of dimension in this field requires more focus and concentration.

Overall it was a positive experience. I expected a little bit more by assignements and by hands-on work in general. I have passed all the grades, but I am still confused about the functioning of tensorflow and why sometimes the assignements stick to Keras while sometimes use tensorflow

Andrew Ng covers relevant and current topics on DeepLearning community, autonomous driving, face recognition and convolutional neural networks. Challenging assignments, and well-balanced quizzes.

Could present the hyperlinks to DeepLearning whitepapers and articles as a course resource.

It is great to play with state-of-the-art neural network algorithms and architectures, but it is a shame that the programming assignments did not involve training/optimization, even on small datasets, or pretrained one (in the case of hyperparameters tuning).

Otherwise, great course !

The lecture is very good, and provided the most recent development in this field.

But the homework, most of time, I am searching the forum try to understand the tensor flow, also the errors from the grading system still exists from time to time.

Overall, this is a very good course.

very intuitive and guided. encourages students not to be intimidated by research papers. promotes open-source software and learning weights. I would've given it 5 stars if it used the latest version of Tensorflow (it seems to be old since it is still not using the eager execution)

The 2nd coding assignment in this course has a bug, in triplet_cost, the expected output is the correct answer but when my answer matches it I cannot pass, the previous versions (v2) give wrong expected answer and in v3 in order to pass I have to match on that. Please correct it.

This course is awesome, just like the other course, but this course required more perseverance and more understanding and more hard work.

Take this course should only be taken if you are thoroughly prepared for it. The previous 3 courses

are quite important for the course.

The course provides a good overview on the most famous techniques for CNN. However, there are several errors in the assignments not solved yet that make you to loose a lot of time. Moreover, it is very hard to do the latter if you do not take all the courses in sequence.

The theory behind convolutional neural networks is very well explained in the video's. Also the programming assigments help you understand CNN a lot better. However, before this course you should have had a course in tensorflow / python. Because the exact syntax is hard.

very helpful and interesting! the only drawback in my opinion is that one who's not trained in using tensor flow can have hard times in figuring out what's happening and what should be done.

I would say knowing TF it's a prerequirement to fully understand everything.

some duplication during the video, and some differences between the course work and the assignment which makes it difficult. Particularly struggled during coding phases as it assumed a greater level of python \ tensorflow skills than outlined in the pre-requisites

This course was definitely harder than the ones before it in this specialization so I know I will need to review this a few times more. But overall, a great intro and insight into how this very sophisticated image recognition and generation scenarios work. Thanks!

It was an all-encomaping course and thanks a lot for providing us with such quality tasks and materials. Just a start for those of us who are trying to figure out what they want and also those who are planning to get choose what topic to dig deeper in the future.

Some things seemed a bit unfinished, some of the videos weren't edited properly. On several occasions Andrew said the same thing over again, like a retake. The code was also somewhat buggy, but it seems they're on it.

The course in itself and it's content is great!

Great course on the theory side - well explained.

Programming side is not trivial unless you have in depth relevant experience in these types of apps. Programming support if very weak (unrealistic) - but there are treasures in the forum (good search capabilities).

Another great course by Andrew Ng. There are quite a few bits in the video that were clearly meant to be edited out, and also some quiz and homework grading have errors. But nothing can be perfect, and I really don't know any other places to learn Neronet online.

Its one of the tough course of this specialisation. Algorithms will not be covered much in depth however implementation details are very clearly explained. You can of course read the in-depth details from the references of the research paper given in the slides.