Couple of suggestions: 1) fix the artwork via neural networks assignment as there is a bug in your code 2) With the lectures I would suggest that you do a summary explanation of how the whole process works (all steps and motivation - a review) at the end of each group of lectures (one for artwork one for face recognition)

I think this course lacks hands-on experience, and for that I think it should improve the labs, people would learn much more from a youtube-kind tutorial than from these "#your code starts here / #your code ends here"-labs.

The quiz interface is horrible, i need to refresh my browser many times before it loads properly.

The theory is very good but the exercise part is not good enough for me (For example in the Face Recognition exercise, I'd like to build (even a simple model) and train the triplet loss function... However, all that I can do is only loaded a trained model and then apply some simple similarity measure on encoding vector)

There is major room for improvement on the automatic grader, under some particular cases, the answers are correct but the grader will give you zero with no feedback (in my case, I made a mistake with declaring local variables as global in the code). This issue is quite frustrating, the forums helped a lot though.

Great class and amazing assignments. I really enjoyed learning about CNNs, YOLO, and Neural Style Transfer.

Errors with submitting the assignments, particularly weeks 2 & 4 took away considerably from the overall satisfaction with the course.

Thank you once again for providing a rich learning environment. :)

Programming assignments were not to the level of the prior courses in the series. Should have more illustration of using Keras/Tensorflow. Assignments either were too spoon fed or there was too little reference information whereas prior courses had a good balance. Many of the keras links are dead.

Theoretical content was very informative and high quality. However, some problems with the programming assigments were annoying. For instance, for the last programming assigment some weights would not load and it was necessary to go get the weights from the github repository of some other person.

The week 1 and 2 were perfect, then week3,4 had some issues with the lectures- Andrew sir was repeating some parts and the problems/corrections in the slides.Also the week3 object detection was tough n the hints were not enough, with the errors in the assignment submission costing me a day

Great content, veerything is clear and concise. Only downside is the grading of the exercises, which sometimes requires you to use a very specific syntax (even though another syntax gives the exact same results) and causes hours of painful debugging and reading through the forums.

Model implementation is abstracted in many exercises. Many helper functions are created to just make things work. TensorFlow feels a little foreign still, not enough of an overview. Higher level APIs like Keras and/or PyTorch might do better here instead of mixing in TF randomly

The last two weeks sometimes bored me and sometimes I had hard time in doing the assignments. The intuition behin object detection/face recognition and neural style transfer are well explained, but some more details for understaing how these models work is missing in my opinion.

The lecture content is good but the programming exercises are not explained well. Quite often you are left on your own to go through Keras and TensorFlow documentation. So, don't expect much help in learning how to implement the theoretical ideas explained in lectures.

The course gives a high level understanding of CNN's, which is good but missing details. The content is okay. However, Andrew sounds very monotone and I happend to lose me focus very quickly in that case. Also, the many errors and corrections are confusing sometimes.

The lectures are very good. The programming assignments are sometimes infuriating and do not add to an understanding of the subject at hand. More can be done to explain the Tensorflow and Keras code. Also complete code explained line by line would be VERY helpful.

I learnt a lot in this course, but i have the feeling that my knowledge is still very shallow specially when it comes to convolutional neural network design, i cannot tell pros and cons of each design and how to come up with new design that meets my use case.

Thank you for your great course, but, to me, it has a great problem. It proposed the general theory of ConvNet and then explained some applications on ConvNet. there was nothing in between, i think it could be better if ConvNets were explained more deeply.

The course is really good, but the assignment grader is a disaster. I spent days and nights reverse-engineering the expected codes, read the forums, only to pass the course before subscription expires, and this is certainly a very disappointing experience.

I wanted to give five star; however, I could not. The function "model_nn" in Week-4. assignment -1 has been very poorly designed/ poorly explained. When I searched the forum, there are numerous questions on the same topic; but,, there was helpful hint.

A great course, but I can't give it 5 stars... There's just too many broken assignments, the videos are barely edited, staff completely ignores discussion forums, and it generally feels a little unpolished. I'm sure this will improve in the future.

Videos are great and give a strong understanding of the concepts, but the programming exercises are underwhelming. I don't particularly feel confident about the hands-on understanding of the concepts after complete the somewhat shallow exercises.

Improve the quality of vedio please. there are too much repeats that could be easily avoided, it much worse than the first two courses, not about the centent, but the vedio itself, is your workers seriously correct the probleme of vedio??????

The course was great but only the first two weeks were sufficient for me as a Mechanical Engineer. I am not really interested in localisation and face recognition. Also, high time that you should update to Tensorflow 2.0 for your exercises.

Great course, but assignments are from trivial to insane.

I would like to learn the ideas, but spending like 10 hours "learning" names and grammar of some packages, that will probably be outdated in next few years, is nothing fun for me.

What I really liked about the course was the actuality of the paper. However, I would have thought it absolutely necessary to explain the BackProp for CNNs. Also the grader problems in the last assignment force me to subtract two stars.

Face recognition notebook has a bug, I passed the grader but the function triplet_loss returned the wrong value in the notebook. Several other people have had this problem despite the fact that the notebook was supposed to be updated.