This course was of the same high quality and detail of the first three, but the last project was difficult due to bugs in the python code. This definitely needs to be fixed please.

The grader didn't work correctly on the final assignment. Course content was both really interesting and well explained though. Thanks very much for putting together these courses!

The content and lectures are still great. But the assignments and graders have a few bugs. I guess it's because of the insufficient preparing time. However, I did learn what I need

The lectures were excellent, but many of the exercises and quizzes had "issues" that wasted a lot of time. You need to spend a LOT of time on the internal forums for this course.

Great content and superb videos. The assignments are becoming exponentially more challenging if one is not familiar with Tensorflow as the intro on that framework is rather short.

great lectures.

yet there are some places to improve :

1) the assignment are harder (which could be good) yet they are hard by being less clear

2) there are some bugs in the grader

Good content, I learned a lot. Just one thing, the programing assignments need to be updated to tf > 2.0. Currently running 1.2.1 (released on Jun 30, 2017) in week2 and week4.

I wish to rate this 4.5 out of 5. Some descriptions (eg: YOLO alogrithm) are less elaborate or may be I found it difficult to understand. Overall, it is a very good course. :-)

Good course but lots of technical issues with grading of programming assignments. Did not see these issues in previous courses. Maybe more background on tensorflow is needed.

A bit difficult for people who don't have too much coding experience, especially those who are not familiar with using open-source packages, reading documentation by their own

It would be more helpful and beneficial if there is a deeper explanation for the network architectures involving the forward propagation, backward propagation and optimization

Another very good course, only marred by the problems with running the final 'Happy House' face recognition notebook of week 4: the values did not seem to load into the model.

The course is good but the grades has severa errors - I spent a lot of time to understand what was wrong with my submissions - and was grade problem - consult always de forum.

the course was good but the implementation of the code for style transfer could've been better it was a bit confusing apart from that the course was good and well constructed

Amazing course, providing a lot of insights in state of the art technology. One advantage might be to provide more hand-on in the programming exercises than just.a overview

Very good content and learning value. I would have given it 5 stars were it not for glitches I encountered in the material available including some of the Jupyter notebooks.

Completed some interesting assignments. Concepts are succinctly explained.

I think deploying these solutions in a real-world environment will show how well I have learned it.

The course is great no doubt however, after completing the course I still don't feel confident in writing code in tensorflow and Keras. Need practice to improve my fluency.

Very interesting content. Exercices are getting a bit messy, sometimes it's just not based on what has been learned coding-wise, sometimes there are issues with the grader.

Phenomenal content, but frustrating problems with Jupyter Notebook. A determined programmer will get through those kinds of problems, but it definitely burned up some time.

In content this is a 5 starts course but, there are issues grading the exercises that have being unfixed for quite a while... I hope they'll get fixed for future students.

Nice Course on convolutional Neural Network. They could have used only one framework to complete the course assignments instead of switching between tensorflow and keras.

Very good and well structured lectures. The assignments make you work through some very interesting problems, but there is not a lot room for creativity in the solutions.

I learned a lot. The tensorflow methods are not explained well enough (even if you did the previous courses of this specialization), and it still uses tf1 instead of tf2.

All good, but something is amiss with the submission of the last week's Face recognition notebook. Use manual "Create submission" under "My submission" and then it works.