A very detailed course with lots of nitty gritty. Learned a lot and of course enjoyed it thoroughly.

short and sweet course :)

It is necessary to do great research, coding and grapple with many mistakes in order to make very effective and possible works with Tensorflow. In this course I learned a lot in a short time. If I researched what I learned here for days, I could only put it together. Thanks Laurance and all coursera team in the back.

My favorite part of this course and other courses in this and other TensorFlow specializations offer by Laurence and Deep learnign.AI is the recaps at the beginning of every video; He connects all the videos and concepts together and makes the learner understand where they are and where they're going and why.

75% of the course was good. Many of the topics were very interesting i.e. how default functions work and all. But the last weak was too hard and was not explained well. Again, it was suggested to use slack instead of discussion forums. But the mentors didn't respond to my query. Hence, the course is a good course and worth taking.

W1 & W2 are amazing, W3 & W4 doesn't help much.

week 1 and 2 were great - full of many interesting information week 3 and 4 were quite boring thus hard to struggle and get valuable knowledge

Was good but the programming assignments could be improved with more practice.

Many lectures are called "X code walkthrough" meaning that Laurence reads through the Python code. The problem here is that he does exactly that, not bothering to explain anything, even if it was an important new concept, etc. Considering that the only other type of lectures are those in which Laurence shows slides or "explains" something while quite obviously reading from his screen, I think it would be fair to add the "slide/code walkthrough" descriptor to the course title. I would not mind Laurence reading from his screen (how could I, he even acts like he was really giving a lecture, with the intonation and all) if he was systematic in explaining the newly introduced concepts. This way it simply does not make sense to me. Why bother with recording the lecture videos instead of just publishing the text that is being read out? I do not advise taking this course. It has a nice selection of topics for sure, but TensorFlow lines are just shown, not explained. You will learn about them way better if you simply read it from the official TensorFlow documentation. How can a course like this have such a high Coursera rating is beyond me.

The course is fine but it lacks depth in applications, assignments, as well as resources.

A complete waste of time.

Great Course,

It would have been laborious for me to try to learn about Tensorflow graph mode and Tensorflow distributed training by myself.

The thing was chopped in very small chunks that were very easy to digest.

The best part is that I now have working notebook examples that I can use.

Thanks

A few things had been difficult to follow, and it took some time to get used to autograph implementation and distributed training strategies in TensorFlow. Nonetheless, it had been a wonderful experience to learn from DeepLearning.AI !

The course is very fine , yet I guess in need more in the practical part , even a recorded video as most of the people dont have multiple GPU on a singel device to detect the effect of implementing mirroring strategy

Superb course from DeepLearning.AI. I have learnt in detail techniques regarding custom training and in depth details of distributed training and how it will effect the performance of our project through GPU & TPU.

Another great course by Moroney sir. Loved how TF can be used to train models using different strategies. A great intro to the deep applications of TensorFlow

5 stars for excellent videos, contents and code walkthrough. Insipired me to learn more and experiment on distributed training and custom training loop.

Detail and easy to understand. I really recommend this specialization to improve training process using distributed training and strategy

It was helpful to learn the details of the optimization by using GradientTape and manually updating the parameters for every iteration.

great to learn things about writing custom training loops, and distributed training of deep learning models.