Comentários e feedback de alunos de Sequence Models da instituição deeplearning.ai

Overall a great course, thanks to Andrew NG for his great explanations. But a very bad support, I faced many issues in submitting the assignment due to technical issues (notebook not saving) but no dedicated resource to help me. I spend lot of time in resolving my self.

Feels again like authors tried to put everything into just a couple of weeks, thus the course turned out to be messy with lots of details hidden. Even though there was a lot to learn, I am still not sure if I understand correctly how to build a simple sequence model.

Really good to understand the basics, however, it doesn't use the latest TF2 and the exercises are either trivial because too much pre-worked, or too hard because it doesn't give the information required to succeed.

This course really needs to be updated.

Programming assignment instructions are not well written (clear), and as a result it is easy to get stuck on something of little relevance to deep learning. Also, I would suggest that you make the lecture notes in written format available.

The lectures cover the basic design of the models but don't help teaching you how to actually use them. I learned more by reading blogs to get the programming assignments to work then this course.

Lectures/Videos were excellent, the assignments were very poor (loads of errors in the code not corrected over 7 months since the course went live)

The course videos were very lengthy and difficult to follow. Many topics discussed in course video were not part of programming assignment

Good introduction to the concepts. Really poor quality videos and exercises. Very frustrating when working on the assignments.

Good lectures. Programming assignments are useless fill-in-the-blank exercises, you don't really learn much from them.

The programming assignments really are like pulling teeth. There's not really enough guidance leading up to them.

The course videos and the programming assignments were lacking. And there was no support in the forums.

Spent more time trying to work around a buggy grader than learning the underlying concepts.

Too hard to understand compared to the previous coursed in this specification.

Was very difficult. please add more expiation of mathematical equations.

Little unsatisfied with the final part of the specialization.

Lack of Keras training made it often hard to do layer code.

This course needs more instruction on Keras.

Poor explanation for alot of things

boring and uninformative

Too difficult.

Needs work.

I feel sad.

sucks

I have no idea how we're supposed to walk out of these courses with the knowledge of how to build a neural network. The practice exercises are a joke. It's a bunch of functions taken out of context, with "instructions" on how to complete each. I don't understand how to do any of it, and I passed all the quizzes.

This specialization gets good reviews because people love Andrew, and although I'm sure he's a great guy, these courses provide no real practical information on how to build neural networks from scratch. I don't even know where to begin, and at this point I'm just copying solutions from the internet to complete the projects so I can just get my completion certificate.

I only recommend taking this from a theoretical perspective. If you're looking to get started with deep learning from a practical standpoint, look elsewhere. This isn't worth it.

Oversimplifying material, such as not showing any math foundations and proofs, does not lead to an intrinsic understanding of the material as well as fill-the-gap assignments do not enhance comprehension.

I understand that the course is intended for the broad audience but will one be able to implement those Keras and TensorFlow algorithms on a moderately complex problem, which is the ultimate goal of these courses? Highly doubt it because the code is pre-written for students and step-by-step guide is provided. In my opinion, one could go straight to assignments and induct / deduct the answers.