This felt like a glorified tutorial for TensorFlow/Keras. I expected more in-depth treatment of the material. E.g. covering more ground (regularization wasn't mentioned at all), or going into more depth on the machine learning theory (why are we using this activation function, this loss, or this optimiser) or practical tips (e.g. discussions of network design) or the tools we are using (e.g. what exactly is TensorFlow, what is Keras, how do they relate to each other, how do they work under the hood).

I also raised some issues and PRs on the github repo for the worksheets to correct problems in some of the worksheets, but these were not responded to by the time I had finished the course over a week later, despite the low volume of issues and PRs on that repo.

I paid for the course upon getting to the first quiz so that I could have my answers graded, but I don't feel that I got my money's worth.

I was expecting a more elaborate content and graded programming exercises for those who pay for the course. Instead the major part of the content is based on the FREE tutorials available on TensorFlow website. Besides, the course is split to 4 weeks but the complexity of the content does not fit to the announced period. It can be easily completed within few days or less by any avearage CV researcher. It is populated with a bunch of very short videos with 'surface scratching' explanations. A targeted learner seems to be an undegrad. Also, some important concepts of TensorFlow were not explained (ex: what is tensor? etc.) Overall, I am disappointed and consider this a waste of money. eventhough, it can be useful as an audit for those who need some structure in self-education as it provides the sequence of tasks and some quizes as many other courses by Coursera.

I like the CoLab Intro, and basics of Keras. But I think the 1st course are a bit too basic for someone who took the 5 courses Deep Learning Specialization. And it is too fast for someone who had not taken the DL courses. I would like the instructors to go over these topics in future courses:

1) TensorBoard and how to debug a faulty model

2) TF 2.0 features (Eager execution, etc)

3) hands on example on how to fix the model if validation accuracy is much worse than training accuracy

4) LSTM models

5) how to productionize the model for real life use, like TF edge or TF.js

Far too basic - almost nothing about Tensorflow - graphs, tensors, estimators, tf.data, etc. The course is a bunch of pre-built examples using the Keras APIs to hit run on in Google collab. This is a very gentle introduction to Keras APIs, with no discussion of what's actually happening with Tensorflow.

I've finished the whole specialization in one week with 100% on all quizzes. I returned to its first course to rate it one star. I would give it 4 stars if the whole specialization was put in one free course "Shallow intro to TensorFlow with non-working and uninspiring examples". The intros with A.Ng are promising, big words and ideas, but the videos and examples are simplistic and don't deliver. There is no much work to do on this specialization, and I developed skepticism and mistrust toward the instructor. He doesn't come across as an experienced ML/TF practitioner, and some of his discussions are intentionally vague. It's incomparable in depth and content to Deep Learning specialization from the same deeplearning.ai. It also dilutes the value of Coursera's Specialization. Please redo.

Really disappointed. Please, give me back my money!

This is not a course of Tensorflow but one page manual from Keras tutorial website. I hoped to learn the Tensorflow but have seen only some very simple examples with API wrapper for simple tasks and passed the course in a day. There was almost nothing from the Title of the course.

The course should be free. There is nothing for $49 in it (except really cool teachers)

Finally, wasted my weekend and 40 euros to finish this specialization. I really dont know the target audience of this specialization. If you have no background of deep learning, going through some code snippets without any explanation wont help you at all. you can't know anything behind it. If you already have some knowledge, you will find nothing new and more in this course. 1) The materials are so shallow and without any depth, just reading the slides and codes with errors. Only some high-level Keras APIs are covered. The official Tensorflow tutorial is much better. 2) The test questions are of no value at all, it cant test any your understanding whether about deep learning or the tool Tensorflow. The assignments are poorly designed, the answers contain errors. 3) I strongly doubt the instructor, I think he does not have much ML experience. Please don't waste your money and time on this specialization. If you want to learn deep learning, go to cs230; cs231n for computer vision; cs224n and cs224u for NLP; cs20 for Tensorflow.

Often coursera courses are a bit easy / superficial. This course is a bit too much so. There's just not enough meat-on-the-bone for my liking. The instructor came across very well, the material is polished and professional, there's just not enough material for me to think of this as being a course. It only takes about an hour to do each weeks' material; the programming examples provide almost no challenge. As a first course on the topic this might be okay, but if you've done anything in this space before then it will be too easy. The course also uses the highest-level TensorFlow APIs; in a sense I wonder if this is really about TensorFlow when that same API is usable with other frameworks. I hope the rest of the specialisation has more detail.

Good intro course, but google colab assignments need to be improved. And submitting a jupyter notebook was much more easier, why would I want to login to my google account to be a part of this course?

Great course to get started with building Convolutional Neural Networks in Keras for building Image Classifiers. This is probably the best way to get beginners into Deep Learning for Computer Vision.

This course barely touches what Tensorflow is and how it works, it is just a very basic introduction to Keras, and most of the time is spent explaining machine learning concepts rather than Tensorflow ones.

Too basic. I took it to learn about tensorflow not deeplearning and it's like just a bit of each which is useless at the end. It's just the 'Hello world' example in TF.

Even if this course it is an introduction to TensoFlow, it is too easy. Good resources and good notebooks, the lectures are not bad and well explained, but the examination part is too soft.

The course demystified simple computer vision classification use-cases by leveraging TensorFlow. This is a great follow-on course to Andrew Ng's 11-week Stanford Machine Learning course.

Nice course with some flaws. It’s a course in Keras with Tensorflow under the hood but you won’t see it. It’s great it’s Keras, however the title of the course is misleading. The videos are only a few minutes per week. Mostly it’s self-study on Google-Colaboratory. If you have no clue about Python or Machine Learning you might quickly be overwhelmed by the coding involved. If you have knowledge in Deep Learning you can earn this certificate within a few hours just by answering the (rather simple) quizzes even without watching the videos because the programming assignments are not graded. The course doesn’t have the depth of the Deep Learning Specialization by Andrew Ng but Keras is a great Deep Learning Library

An awesome practical course that helps me to start creating my first neural networks using keras in such great methods, the instructor is very good at delivering the knowledge he has

.

It was a waste of time. The most part is old with tensorflow 1.14 and the code in the course didnt fit with the coursera Lab environment. Also I wanted to do the Tensorflow Engineer Course and thought I doing this in my test time and get a certificate for this...but it brings me to this page and I just get the introduction. It told me I get the full Course material...

Very good and concise introduction to tensor flow. Important parameters to methods are explained so they are understood and no longer mysterious. The course starts with a very simple "Hello World" type model and builds to a more complex CNN. During this evolution, the code becomes necessarily more complex but the additional methods are explained as they are introduced.

I had a great time going through this course. I had a lot of challenges but which made me love the course even more. While I'm excited to start the next course, would still like to go through certain areas of this course to get a better grasp on some areas. I'm grateful to Cousera for such a learning opportunity!

There is quite some bugs in the example exercise code. Hope they can improve this in the future. Despite that, the content of the course is well designed and is a good start for beginners in TF