Jan 09, 2018
Thank you very much for making this course available on Coursera, I cannot agree more the knowledge of Mr Venkat. This is a great way to help people to get started with Google Machine Learning.
Sep 08, 2018
A very good course on TensorFlow, ML and Google MLE on GCP.\n\nThe Labs are self contained and the problems proposed are very challenging. I learned a lot on this course.\n\nThank you!
von Anna K•
Jan 19, 2019
It was rather confusing than helpful. A great overhead of doing the same configuration stuff for the first 4 labs: They could be combined to one greater Lab. The last 7th Lab was too complicated, I lost connection to the notebook all the time and had not enough time for understanding where and how I should run the predefined commands. I rather submitted the commands without thinking how, where and why, because I did not have time to think: The lab is too big for the short period of time. I would appreciate to have more time for understanding the commands (for example for ml engine and dataflow).
von Joose R•
Apr 26, 2019
I was a bit disappointed at this course. It is after all data engineering track course but it seems like it's data science for those with ample data engineering skills. The course focuses more on basic ML concepts and less on the Google Cloud Platform specific boilerplate code. The things that are specific to Google Cloud are just viewed cursorily and I didn't really get much insight to them. Other issues with the course are that it's basically one big deprecation warning. Apache Beam cannot be run on Python 3 but Python 2 is about to deprecate. Furthermore, ML Engine is renamed to AI Platform but the course material hasn't applied that change yet. Instead it speaks about ML Engine. Additionally, distributed training isn't covered very well but it's marketed that Google Cloud does distributed training on its own. So basically after this course I learned that the services used in this course are probably not used by Google because everything is deprecated and about to become obsolete and the services are quite cumbersome to use.
von Michał J•
Aug 28, 2018
To be honest this whole specialization is a scam. There is promise of preparing to Data Engineer Cert, but the course just skims most of the topics, missing a lot others entirely. Please, at least have decency not to advertise it like this and remove it from official google certification page.
von Joel G•
May 26, 2019
I wasted to much time every lab setting up the environment with not enough time to go through and actually understand complicated code.
von Alexandre V G•
Sep 14, 2018
Lax is horrible. He's going around and around and around and around and around... and repeats and repeats and repeats.. mostly useless stuff. The course is Boooring... Not much about the GCP. Mostly 'intro to machine learning'
May 02, 2019
lab 7 is rubish
von Mussie N•
Feb 18, 2019
This is a great introductory course to Tensorflow. I prefer if you can explain some of the concepts using animation but that may be just my preference. However, I would love if you can think about it every time you explain a more complex concepts.
von Naveen B•
Oct 08, 2017
The course appeared to be put together poorly by pulling from many different sources. Therefore it lacks coherence. There are many videos that are under 1 minute long and just there to make a statement or two or just to repeat the last video. Some videos end before instructor could complete his statement. Some videos are recordings from a classroom setting and makes it hard to follow. Also would've appreciated some support from the course team for questions posted on the discussion forum. I think topics such as machine learning in GCP deserves a bit more respect. This course needs to be planned from scratch with MOOC and Coursera format in mind.
von Miguel P F A F•
Dec 15, 2019
I liked the materials and the instructor. However, I think this course would be better in 2 or 3 weeks, instead of just one. The way the information was compressed, both in lectures and in labs, makes it difficult for the students to retain important information in way that would enable the student to be autonomous and create own projects. The nature of GCP and MLE is complex with lots of connections between modules, inter-dependencies, etc. Lots of aspects of the code were not given the necessary focus and attention. Maybe with a longer course it would be possible to focus in all details, in a step by step way.
von Graeme C•
Oct 23, 2019
A number of videos cut off mid-sentence.
The video flow was poor.
Some videos were too long, some were too short.
The code for TensorFlow was absolutely blown through, and then time focused on other stuff.
Most of the time in the labs was spent waiting for SSH keys to propagate - 3 of the 4 labs in the TF module(I think that's the one) should be combined and made into 2 hours total instead of 1.5 hours each because they are all extremely related and yet had separate, extremely long setup times.
von Mark D•
Aug 12, 2017
Should be a much longer course with more hands on programming instead of just reviewing what is written. So many videos are 30s or less, videos are cut off, or following video is incorrect. In at least one instance a video was duplicated out of order.
von Aliaksei K•
Sep 30, 2017
Thank you very much!
Even though there are some points that may be improved, the course itself is a very useful source of learning ML in Tensorflow. The examples are explanatory enough, the instructor highlights the key points and draws students' attention to them multiple times so that the meaning is understood. Besides, the way you do exercises in Datalab is simple and in the same time interactive and full of information.
Points I believe to be improved:
- some videos with slides are too small. In some cases, you wait more for the Coursera engine to load than the length of the video;
- some videos can be split in a better way so that either the endings are not cut out or several parts of them are not repeated;
- please either split the Google Lab into parts or put a more noticeable note (e.g. with a bigger font or before the link) so that the lab's exercises are done along with the course flow;
- I did not notice a link to the 2d lab;
- one of the first videos in Module 3 contained questions right in the video; I believe it is a good idea to put them in the other ones so that students pay more attention to the information and actually have no choice but to think over the problem;
- I believe questions from the audience can be put in a textual format in videos rather than scrolling the transcripts to see them.
von Abhinav V D•
Aug 20, 2017
The online MOOC Course which is having the title Server less Machine Learning with Tensor flow on Google Cloud Platform was successfully completed with the final grade of 100.00% and was found to be very much helpful and exciting as it expanded the horizons of my knowledge regarding the broad domain of Machine Learning and Pattern Recognition and also contributed in building up a strong platform about the various basic skills and techniques that are used to solve a problem that is related to the field of computer vision and is user oriented in nature.
Aug 17, 2019
Really awesome and a very well-curated course.
Long courses may make you loose focus. Small courses may not have enough information. But this is the really awesome course curated with only the information required. Really I enjoyed and got the required knowledge to start to google cloud ML and machine learning.
I suggest that understand machine learning concepts clearly before getting into the course. So that, we can follow this session along with the way very easily.
von Muhammad N B•
Aug 25, 2019
Loved the course outline and learned a great deal. Not only did I understand ML model creation lifecycle, how to improve accuracy significantly using feature eng, hyperparam tuning and Big Data but also acquired knowledge of state-of the-art GCP tools such as Cloud ML Engine that can help train and serve these models at scale with ease.
Big Thanks and round of applause for Lak and all those who helped in putting this together.
von Sachin G•
Sep 28, 2017
Excellent course for learning to use Tensorflow and deploying your model on Cloud MLE. Definitely recommended even if you are new to entire Model creation process and using Phython etc.
Some minor ambiguities - e.g. a few times references to slide pictures in the talk was not clear where exactly in the picture the speaker is referring to, when he was saying 'Here'. Still it's a complete 5 star course
von Amadeus M•
Jul 22, 2017
Great introduction on how to do scalable machine learning on Google Cloud. Easy to follow as long as you have some basic understanding of Python and database queries. The real learning curve will of course only happen once you have to actually apply this to your own datasets, but the course gives you a good overview of what Google Cloud and TensorFlow can do for you.
von Michael O•
Mar 17, 2018
The course covered some very practical aspects of feature engineering as well as big data concerns that were new learning for me and provided some good intuitions about how to design good models. Also, the set of labs provides a very nice set of example code to work from on future efforts - providing a nice map for exploration of other areas of the GCP ML and APIS.
von Chris R•
Nov 17, 2018
I appreciate the difficult problem solved during this course. The student uses many of the tools available but also gains insights on machine learning concepts in the process by working through a comlete solution. Lak is a great teacher and while it's not my first exposure to mahcine learning, this course helped me solidify concepts I've learned in others.
von Dushyant S R•
Oct 01, 2019
one of the better course , what is not good is support that we get for resolving LAB issues , any isse with LAB , standard answer from QuickLabs would be "the labs work fine when we do it", which kind of defeats the purpose of having a support.
Double whammy Coursera Support , I guess cannot do much apart from saying "we cant do much , its quicklabs"
von Arvind K•
Feb 14, 2019
only Lab7 has troubled me due to incorrect code for the installation. It should be:pip install --upgrade pip
conda update -y -n base -c defaults conda
source activate py2env
pip uninstall -y google-cloud-dataflow
conda install pytz==2018.4
pip install google-cloud-dataflow
pip install apache-beam==2.5.0
von Abilio R D•
Dec 19, 2019
Very good course, with lots of good labs and materials. The videos where also very good , and the review demo after each lab complement the information and help one to understand better the decisions made in several steps of Tensorflow ML programming. Congrats for the authors and big tank you!
von Parag G•
Aug 17, 2018
It was a nice journey through machine learning paradigm from very basic to complex machine learning models. Built good understanding of tensorflow and how cloud mle works with tensorflow. Then it described a good understanding of how cloud mle can be used in live system.
von david m•
Aug 04, 2017
Excellent course. Quite a lot to take in at times especially if you don't have a maths type background. The lessons are good and paced well. This course seemed significantly longer than the others in the series, probably as it was out of my comfort zone.
von Edgar H•
Dec 19, 2017
very informative. contained the right level of detail for a course of this length/duration. having been exposed to the core / basics of ML on GCP and having been shown the overall Tensorflow-based ML pipeline structure makes further exploration feasible