Comentários e feedback de alunos de Structuring Machine Learning Projects da instituição deeplearning.ai

This course is just a set of (perhaps useful) advice on how to make decisions when working on a project, not a course on techniques or how to actually do things. There are no programming assignments as in the other courses of the specialization, just some "decision making simulators". I learned more and enjoyed more the other courses. It feels like all these advice could be given as part of the other courses. (But perhaps I am much more technically inclined.)

This part did not interest me much because I find that it does not go into detail and concretely I did not learn anything useful. Indeed we have plenty of examples that teach us what to face in a situation but in the end if we are a beginner we simply do not know how to do ... I find that it is + a documentary that Classes.

I am hard on my scoring of this 3rd part but I strongly recommend to follow the first 2 parts which go into detail.

Although the content may be useful for Deep Learning researches/practitioners. I think there is no need to have a stand-alone course but rather include these guidelines or best practices in the first two courses of this specialization. Some of the concepts are as well repeated. There are no programming assignments or any other way to 'visualize'/'practice' the ideas mentioned here.

The most exciting part of the course as others in the series is the interviews that Andrew does with deep learning researchers. I thought I would learn more about how to structure actual machine learning projects from a software perspective and how I would incorporate them to real products. I felt the videos for this course were too long and cover somehow basic common sense.

the course doesnt have any programming assignments. I feel that these two weeks should have been added/combined with first 2 courses. The knowledge that is provided is useful, but it is mainly useful once you are an expert at building neural networks and models. I feel that this course should have been the last course in the series instead of the 3rd course

Just a few videos without any programming excercise or a bunch of rather broad statements that are not really tried out in programming examples are not really worth the money and more importantly the time. The first two courses are good, this is definitely a drop in terms of quality. This one needs more meat on the bone.

This had the potential to be a very good course, but fell far short, in my assessment. It probably should have been rolled into the previous course as a couple of additional lectures.

Either that, or it should be expanded greatly, with more practical exercises to solidify the concepts taught.

Doesn't look like it was checked for quality control (e.g. Videos with bad takes), Ng rambles sometimes so that it seems as if he is filling time, there are no knowledge checks. This course wasn't ready. Case study flight simulators are good, but poorly introduced.

The teacher seemed to not have a clear idea of what all he needed to say in the video and verbally flailed somewhat, and many times seemed to be adding things purely for the sake of padding video length. Don't waste my time. The content was mildly useful.

I need a lot more practice than is offered here. I would also strongly prefer if the instruction followed some of the best practice laid out in books such as "How Learning Works" because I have difficulty following the instructor's line of reasoning.

Answers in the multiple choice seems incomplete for me, I don't necessary agree with the answers presented unless more detail on some context (that I don't think we should assume) is present for the questions.

This course is good, but here is so few practice. This is hard to understand without practice. Looks like I didn't understand many problems in this course. Need more explanation, more samples, more practice.

Zero programming assignments, but simple quizzes that will make whatever you just learned as fleeting as the morning dew on a hot summer's day. Too bad, because otherwise the material is quite interesting.

Although important, i think this should be the last course in the specialisation as it covers the harder parts of handling various errors and their causes before knowing how these models are trained

the best course in so far, not that much theory but a lot of "insides" from the field. However, still no practice, Im studying for 3 month and still have no idea how to create a real application.

There's theory, but, without practice and application in my opinion. I did not like it because it seems to be easily forgotten seeing that I did not associate with practical excercises.

Most (if not all) of the information covered in this module was covered, perhaps with a little less depth, in the previous modules. However, it's probably worth repeating.

Poor video editing. Not enough graded material to feel confident that I fully understand the concepts proposed in the lectures. Definite step backwards from courses 1-2.

Could be better in terms of the concept taught. A course I would prefer as the last one in the specialization. Week 2 Material is good but whole course is too slow.

This course had a much less ambitious scope than the previous two courses and I think that the programming assignments are very important to help me learn properly.

Too much information in too little time. Additionally, all information is mostly practical, and having no real exercises makes it hard to remember all the details.

The course should be much shorter (e.g. half a week). The messages are pretty straightforward and could have been passed in one quarter of the time.

The videos are to long and it presents some topics purely hipothetical. You basically spend a couple of hours without developing any useful skill

I still feel a little confused when I have so many options to improve my NN. This course is less clearly taught than the two former to this one!

I found it quite verbose... Could have easily been shrunk and fit inside the other course... Don't think it needs a separate course for this