good

good

good

Good

best

good

good

good

It is short and comprehensive introduction. It could have had a dedicated module on evaluation of the models, with visualizations of target vs predictions and losses. From evaluation of peer-graded assignments I get the impression this is not well understood (ways to do it, meaning of values vs training and epochs). On the other hand peer graded assignment should be more challenging than what is shown throughout the course. So maybe it is enough what was shown throughput the course, as current assignment is a bit more challenging. Otherwise students end up copy pasting materials (which I have seen too often). My problem is more on the concept of evaluation of the assignment and points to be given. Scale is too coarse. And submission request should be less loose - jupyter notebook or python files, not html or pdf files. And some system that is automatically checking for similarities among student's assignments prior to submission would be good to have.

Very good course which gives a good introduction to the field. Don't get intimidated by the math you will see and make sure you understand the workflow. Once you do that you will basically repeat it in which one of the neural network types presented at the course. In a negative not, I missed the intructor elaboring how to identity problems that could be approached by applying DL. But I complemented studies on other documents in the internet and that's ok.

It was a pretty good brief, rapid intro. I frankly was expecting more content on options and explanations, but it covered the very essential basics. The final exercise did ask for students to use tools not gone over in class (a bit of scikit-learn). Since I've used scikit-learn before, this wasn't hard for me, but it may be for a newcomer, and actually isn't needed to meet the goals of the assignment, so I'm not sure why it was there.

The course contents are not in-depth enough. The server for Jupyter notebook running is way too slow. Besides, the peer review homework is not that good, because some people didn't read through the questions carefully enough, and they misunderstood the questions themselves and could not give fair enough grades to peers. If the final assignment can be made to auto-grading one, it would be much better (we can set the same random seed)

I have experience of Deep Learning, so I am able to walk through this Lession quickly. The main focus is on Capstone Project, and I have learned something on it. To be honest, this lession is very elementary. I suggest to introduce more Deep Learning models and approachs in this lesson.

This is a great course. The lectures are boiled down to the essence of neural networks using Keras. I give four stars instead of five stars, because the IBM labs environment that the course uses was quite slow and buggy, so I ended up doing the exercises in Google Colab.

The Course is basic but interesting. I missed an exercise on backpropagation with the same explanatory level that it had for forward propagation. The last activity needs to be reviewed because it is confusing.

Nice overview. The coding exercises could be deeper, and in the second half of the course lose any depth at all. Understandable for such a short course, but still felt like a missed opportunity.

A good course. Could be better if it was explained how to select the optimal number of layers and nodes. This was not covered and explained anywhere. Overall it was good.

I took this course for understanding the TensorFlow properly. Now I am in the situation to understand all the frameworks. Thanks a lot for providing me this free course

Good practical examples for ANN. It could be improved the theoretical part and compare better the architecture of the networks with the algorithms and code for Keras

The course is quite complex for a person who does not have knowledge of algebra, statistics and calculus, the final project was good because it was challenging.

A good introductory course, well suited for beginners looking for general information about neural networks and deep learning, with good practice exercises.

Very good course. If we could have the answers to the projects after submission, that would help a lot. Please see if same if possible. Thanks,

Danen

Good intro to the subject. The instructor could have done examples using other neural networks like RNN and autoencoders. It was a fantastic intro

It is a good insight for someone to know and understand Deep learning. And exams and projects make sure students learn and practice new concepts.

In-depth concept-analysis is required. Good for people who know the theory and want to learn and revise its implementation in Python.