Chevron Left
Back to Browser-based Models with TensorFlow.js

Learner Reviews & Feedback for Browser-based Models with TensorFlow.js by DeepLearning.AI

4.7
stars
692 ratings
163 reviews

About the Course

Bringing a machine learning model into the real world involves a lot more than just modeling. This Specialization will teach you how to navigate various deployment scenarios and use data more effectively to train your model. In this first course, you’ll train and run machine learning models in any browser using TensorFlow.js. You’ll learn techniques for handling data in the browser, and at the end you’ll build a computer vision project that recognizes and classifies objects from a webcam. This Specialization builds upon our TensorFlow in Practice Specialization. If you are new to TensorFlow, we recommend that you take the TensorFlow in Practice Specialization first. To develop a deeper, foundational understanding of how neural networks work, we recommend that you take the Deep Learning Specialization....

Top reviews

EJ
Mar 17, 2021

This course is very practical and interesting.\n\nI enjoyed the excitement I got along the way.\n\nIt was modeled to make you pass as long as you want to pass.\n\nThank you Laurence and Andrew.

AZ
Mar 28, 2020

No doubt, the team of Deeplearning.AI is building best learning resources. We would love to get more and more resources for easy way learning from DeepLearning.AI Team.\n\nThanks to all.

Filter by:

151 - 165 of 165 Reviews for Browser-based Models with TensorFlow.js

By Stéphane D

Jul 12, 2020

Laurence Moronay is really a great teacher and the course is very interesting and pleasant.

I Have removed 2 stars for the time wasted trying to make the examples and exercises provided with the course work :

=> Mobilenet Model version not compliant with the grader

=> A lot of WebGL issues (solved by setting backend parameter to "cpu")

Suggestion : report these issues during training to avoid hours spent on the forum

We can see people who already had these problems months ago and nothing is done to improve things

By Francesco B

Nov 2, 2020

The course is good and interesting. It gives an overall idea on how to embed TF models in js. As in other courses of this specialisation, it is not an in-depth course but rather a fast-forward one: in my opinion this is good if you are not interested that much in these topics, not enough if you want to go deep. Nevertheless, contents are still comprehensible and concepts quite clearly explained. However, one might find more than one difficulties when trying to implement something by themselves.

By Tryggvi E

Apr 12, 2020

Mildly interesting to see this work can be done in JS, but from my viewpoint: Why? I already can do it in Python... I am only stepping through this course on my way to the third and fourth courses in this specialization.

By Chris K

Apr 18, 2020

Quizzes are based on syntax and spelling, which feels like a waste of time. Questions should be more about concepts. Examples are pretty basic.

By Igor M

Jan 3, 2020

Too basic. All exercises are copy paste from the shown examples. All 4 weeks you can complete in just 1.

By Simon O

Jan 19, 2020

Not as good as previous deep learning courses. The exams could have been a bit harder.

By check l

Apr 4, 2021

Explain the accuracy requirements for the assignments

By Abungu B O

Nov 1, 2020

ohh the last assignment on rock paper and scissors

By Stephan S

Dec 23, 2019

A lot of coding and only a few ML/AI concepts.

By jeff s

Apr 19, 2021

The intent to teach DL application in the browser is well scoped. I would suggest more background reading including fundamentals and articles outlining state of the art (challenges and approaches - successful and not), and then additional reference material surrounding related areas such as execution on the edge device. But nothing should excuse coursera or the teachers from technical difficulty experienced by many (dozens of students) with grading. Who is at fault? You wasted my time. And you let dozens of students collide head-on without warning into this problem. You should, at the very least admit to your failure and more importantly compensate me one full course at no charge. Without compensation, i see no reason to trust you. Equally important, Coursera is an off-line teach ing system. I make mistakes. I get so little feedback, your system blocks my learning. There is incredible value in learning what not to do. Again, you need to fix this across all courses - and this goes for all such platforms. Thank you, there is some value in taking a course - we all want more

By Michał O

Jan 17, 2021

Not for the impatient.

3rd and 4th week assignments are quite problematic:

3rd week is heavily dependent on python and libraries versions in user's environment and the course does not point which combination is the one which will be graded correctly. Took me 3 long debugging sessions to finally find the right one.

4th week involves taking had photos to get some gesture recognition. I've spent over an hour with constantly crashing web browser and different cameras to get it right.

Although the skills that can be learned in this course can be really useful, the unnecessary hardships prevent me from recommending it to anyone.

By Jochen R

Dec 17, 2019

it is very exhausting to pass the tests due to hardware and software prblems, though the programming is very easy

By Vitalii K

Dec 16, 2020

A lot of explanation of obvious things. Also, excercises are low quality, with Week 3 and 4 quite hard to pass because of technical issues. Week 3 - need specific versions of the libraries, which are not provided, without which "the model is invalid". Week 4 - quite hard to collect training samples for the model to reach required accuracy, since the app crashes after ~200 examples and wipes them out - you have to start again.

By Musalula S

May 2, 2020

The course content is very good but the instructions on how to install Tensorflow 2.0 and Tensorflow.js in python 3 are not clear.

By szymelfenig

Oct 11, 2020

week 4 submission....