Very good course.

Good knowledge.

Good lecture...

very practical

Challenging

mid

The content of this course is actually quite good. Videos are clear. There is no material overload like in the Course 3 of this specialization.

The bad thing is the same as for all the previous courses: the absolute absence of any graded coding practice. All of the "shift+enter Labs" are hilariously inadequate to help retain anything you've learned during lectures. It's like learning PyTorch by watching how your neighbor is writing all of the code. A complete waste of time.

In addition, the quality of the quizzes has also degraded quite a bit when compared to the very previous course. The idea that anyone would be challenged by a quiz with only 2 questions, each with a multiple choice out of 2 potential answers, one of which is always nonsensical, is just laughable. Makes you wonder what kind of audience they made this course for, as the absolute majority of the quizzes are of this retarded type.

The content of the course is interesting and light and allows for a good basic understanding of the topic. However, the presentation is not good. The automated narration in the videos is often weird and repetitive. The visual presentation is also not great, the colors and diagrams help very little with the understanding of what's happening in the routines (except for the skecthes of the neural networks themselves, which are pretty good). The lab notebooks are riddled with typos. The course would be improved by having a more detailed discussion of what's going on in the code before practice in the labs. Otherwise, I recommend the course for anyone who is starting in the field.

This course was not to the same standard as some others I've taken on Coursera. I think the concepts would have been very hard to follow if I hadn't already taken the Deep Learning specialization, so it isn't a great conceptual introduction to Deep Learning. That said, it also doesn't deeply explore the nuances of the PyTorch library, or give very much guidance on best practices or how it differs from other popular frameworks like Keras/TensorFlow. The quiz questions are fairly shallow (and often frustratingly ambiguous). Probably the best part of the class are the ungraded lab assignments.

I like to feel a human is teaching me, but I felt a robot is teaching instead. Also, many parts of the labs are copied from each other, and that's good, but the sentences and comments are forgotten unchanged and they don't belong to that lab. I preferred if the PowerPoints were designed better. Other than that, I can see that the author tried his best to include everything.

The presenter at times goes too fast and once he's finished talking the slide moves forward before there is time to absorb the material. The slides also contain errors. Should be more throughly reviewed. The labs also contain some bugs. The quizes contain some spelling mistakes and some of the quiz questions are unclear.

I have just one thing to say: there is too much packed into a single course. Astonishingly, this course includes 44 notebooks, each of considerable length. I believe it would be more appropriate to split this single course into three or four courses and organize them as a single specialization.

Potentially a good course, but due to the very short videos and complete lack of supporting material (not even the slides of the videos), it's hard to follow. You need to watch the videos over and over again and take notes. Not sure why not even the lecture slides are provided.

The videos were informative, the Jupyter notebooks were also fine, however, when it came to using Watson Studio, there were way too many problems to deal with. Perhaps a more user-friendly approach for the final project/assessment would have been better.

Very theoretical course. You can claim the badge without running any code. The additional honours course consisted of a total of 3 lines of code you had to write. I did not really enjoy this course. It covered a lot of things but was as dry as possible.

It is very fast paced. There are a lot of videos and not enough opportunities to actually reinforce what you have learned in terms of shorts projects. I think at the end of each week, there should be a small project in order to progress.

The lectures and quizzes are too short to provide much value. The material could be better condensed. The labs were useful, although at times, it felt like the same material from a prior video. In general, it was a good overview.

Ok walkthrought of pytorch, a lot of content but slight mismatch between rather basic DS topics and advanced programming skills. Materials need to be reviewed for spelling and grammar, some quiz questions are unclear.

Overall a good introduction to Pytorch and machine learning. But, some of the questions in the assignments have the wrong answers and/or are worded in a confusing way (can be interpreted in different ways).

the content is definitely overloaded.. i am blowing.. felt like i went to college again. if cut some of the content it will be much more user friendly to learn.. for an online class prespective

The topics selected for the course are good, but the implementation is far from perfect. The quizzes seem to be generated by some machine learning algorithm. Expected better from IBM.

The course is quite complete, but it contains to many things already contained in the previous courses within the Specialization. The final honor part could have been much better.

The later labs had little or no practice exercises. The typos/mistakes in the materials were a distraction. It was disappointing to not have a fine tuning lab in the last week.

There were a few errors in the lessons (grammatical usually, occasionally more severe). All in all, I would recommend this course to others.

Ok to show you how to use pytorch, but to learn ML you should really take the famous course everyone is talking about, it's much better.