Avis et commentaires pour d'étudiants pour Browser-based Models with TensorFlow.js par deeplearning.ai

Excellent

Amazing!

kerenen

So cool

Awesome

GREAT!

Thanks

Great!

COOL!

Good

ty

The audio volume of the videos in this course is extremely low (compared to 5 courses I took by Andrew Ng) and make it very hard to listen. I have to max out my speaker to listen to them.

Also, the version requirement tensorflow for the Week 3 assignment needs to be emphasis to avoid wasting everyone's valuable time. Also, many on the forum were suggesting tensorflowjs = 1.2.9 (which incompatible with tensorflow 2.0.0) is the key to complete the assignment misleaded me and wasted many hours.

The course is easy to follow up and is very easy to pass the graded exercises as you only need to copy-paste code from examples. However to understand what you are doing you have to have already knowledge of tensorflow as such. This course is great if you want to make browse-based ML applications and need few (basic) examples.

However, the discussion forum is lacking qualified tutors very much. There are tons of questions without answer. This is something deeplearning.ai should focus on.

The course is exciting and very practical. Regardless of the last assignment in course 4 which is not very clear on what is required and whether the samples you are providing are correct, as some of us have low resolution webcams and are dark from Africa. As compared to the Tensorflow in practice, the deep learning team is building momentum. Overall, good work guys. Excited to go to the next course.

the course was really good and gave me a good start to tensorflow js but there were some points didn't be cleared much like what is the drawbacks of transfer learning as 2 models not only one and what is actually done when we load a model from online json file.

I learned a lot about the bower based machine learning models and Transfer Learning using JavaScript in Tensorflow js. Also learnt using the python model and architecture in java script for save training time and improving the efficiency of the models.

It's a super interesting course but at it end it feels that it's too easy. I think it could be more challenging either on the depth of the konwledges taught or in the complexity of the exercices.

But as a quick introduction is nice.

The contents is much easier than my original thought it would be. But it might be because the personal experience I already had. Recommend course to someone who just finished Deep learning and have not too much coding experience.

I did Laurence's previous course on Tensorflow, and this one is surely the continuation of it. Rather than makiing it complex by using jargon and big words, instructor has done a great job explaining things.

course contents are good and explained very well with one problem of audio, audio is not clear and pitch is low but I like this course. as a beginner, this course is best.

i like how Laurance give the explanation, but there is something that I think very bad, that is a program assisgment is complicated and takes a lot of time to complete

Good as a introduction for application based machine learning, however, I think the course should add more realistic example of such interesting process

Good basic concepts explanation, but so much copy/paste exercises between them. However, you can learn the main uses and it is a perfect start point.

Everything about the course is good except for the last assignment. Took he several tries to clear it, not so straightforward as other exercises.

it was good, but it heavily depended on knowing html, but it will help with the basics when someone is creating a model for web page or smt