At a couple of important points in the course (e.g. where it is about launching TensorBoard or even more important where it is about deploying the model with ML Engine) the code in the Lab differs substantially from what is shown in the discussion of the lab. This is a little irritating. That aside, I have learned a bunch of new techniques and processes to improve my coding and especially: code more quickly and scalable. Thanks for some really good lessons.

Very useful but I had some problems with lab infrastructure. Options to create buckets wouldn't appear sometimes and I had to open and close google cloud console to make it work sometimes. Regarding the course it was great but there is a lot of boilerplate code and though the steps are simple and clear there is a lot to digest, I will need much more time master this TF/GCP workflow, but anyway this is a great start.

I think a lot of the lab-explanation given in the video following the qwiklab should be in the python notebook; make it a little more illustrative (e.g. architecture diagrams). Also, be a little more generous with the lab time - the last lab was too long (or perhaps change the code to select the faster ML option - standard/TPUs etc. to make the training go faster)

One of the lecturers should improve his English speaking. The course should add more contents, explanations, and exercises for the 3rd part of the course regarding how to scale TF models with CMLE, for example, some bash cmds or some code are confusing, unless this content will be covered more in the following courses.

I could not get my final lab project to work. I have sent the issue to Qwiklabs - I got the following error message:

ls: cannot access '/home/jupyter/training-data-analyst/courses/machine_learning/deepdive/03_tensorflow/labs/taxi_trained/export/exporter/': No such file or directory

I enjoyed this course a lot. If I could modify anything, I would adjust the content and pace of the third week. The videos are relatively simple to understand and well-explained while the final lab feels a lot harder with a lot of unknown command to execute.

hands on demonstrations were good. More in depth explanation can be done fro some of the codes including the part in which data fatching from the json file was explained, and the process to be followed in the gcp to make the model and deploy it.

Excellent course in the capabilities of tensorflow, the course material and data-lab examples are super useful and provide a good overview of how to implement tensorflow models locally and in the cloud with high-quality practices.

Challenge problems at the end of each assignment are really good, however, there should be videos showing how the instructors would solve them, I would be fine watching 30 min videos describing the solutions. Nice course!

Intro to TF should have packed with more fundamental concepts around TF alongside existing topics covered. Moreover, some of the code needs either further explanation or references to understand what a given code is for.

There were some technical problems. Some of the exercises could be described in more detail with TODOs.

Nevertheless I very much enjoyed the course. Quite an amount of material. Challenging tasks. You can learn a lot.

Was expecting a bit more around tensorflow basic concepts. Coverage was too much from basic to production level deployment. Was expecting a bit more hands-on on tensorflow basics and details around deployment.

The tools and methods presented were great. The instructors were also fantastic.

However the coding exercises were lacking in guidance even though the complete solution is given in the video.

This needs some updating - looks like the Tensorboard is no longer accessed in the same way as it was when this course was produced.

Otherwise great! Good challenges! :/

The first week was the best, as it described some of what's going on under the hood. I would have liked much more on these topics and less on specific cloud products

The procedure to connect to the cloud datalab was time consuming to do it every time.

Suggestion : More topics in Core Tensorflow could be added. I enjoyed the course!

Good general TensorFlow introduction. The course focuses on the highest level tf.Estimator whereas there are maybe something interesting low level they don't present.

The course covers quite a few concepts -- TF basics, TF estimator, Google Cloud ML. It would be easier if the material is split into TF and Google Cloud lessons.

Great course as an introduction to TF, however, the labs are not as in depth as I'd have liked. Nonetheless, the course is well executed by the presenters.

A good understanding of bash cmds and a well-digested understanding of the course material is required to perform the labs. Quite challenging.

Good course. Familiarity with gcloud command should be a prereq. Thank you to Coursera and the Google team for putting it together.

Week 1 and 2 were great and quite detailed. Week 3 was too condensed. Overall better than Course 2 of the specialization series.

Not very elaborative in Week 3. Correct me if I am wrong, the 3rd week lab feeds in still in-memory data from a local file.

Good course to understand basics of Tensorflow. Its better to update the course content with Tensorflow 2.x. Thank You.

Good course! Sometimes I was not 100% sure what I was supposed to do. But the solution videos made it clear afterwards.