“TensorFlow Developer Certificate in 2022” — Another great course?
As a follow-up to my previous article about Zero to Mastery’s Machine Learning & Data Science BootCamp, I considered reviewing this course as well. Indeed, when I decided to purchase the Data Science BootCamp, I also took the TensorFlow course — because I had no idea what it was for, but I’ve heard this name so many times I wanted to give it a shot —.
And indeed, it was a good idea. First of all, because Tensorflow is one of the most used libraries for general Deep Learning along with Pytorch, but also because the TensorFlow Developer Certificate is a great certification to kick-start your career in AI ( I haven’t passed it yet, but I’ve read so many articles about people explaining how beneficial this certification has been for them).
And if you look at the Tensorflow Certificate Candidate Handbook, you’ll notice that the course content is exactly what you need to know in order to pass the exam. So let’s review the content.
Course Structure
Deep Learning and TensorFlow fundamentals
It’s always great to learn something new, but what’s the point if you don’t even know why you’re doing this? This section allows you to understand the stakes of making the AI grow, and why Deep Learning is more and more important.
Thus, Daniel (the tutor) gives you both the intuition about why Deep Learning is necessary, and how Deep Learning works from a very high-level view. In addition, you’d learn how to perform basic operations with Tensorflow. As I said in my previous article when I talked about Numpy & Pandas, it might not be the most interesting thing to learn, but it’s crucial if you want to become more comfortable using Tensorflow.
Fundamentals
The course has numerous chapters claiming to be the fundamentals of the X field with TensorFlow. And indeed, Daniel does a great job at explaining these concepts in a very simple and concise way. As for the data science BootCamp, he doesn’t dig deep into the details. Instead, he would spend a bit more time asking how you would solve the current problem by hand, so you can get an intuition about the algorithm used. And when you think about it, you don’t need much more when you’re getting started with Deep Learning: you need a basic understanding of the algorithm you use so you can tune them, but you don’t need to know the exact formulas to be able to use them.
Among these chapters, you can find:
- Neural Network Regression ( how to estimate a value with Neural Networks, i.e. the price of a house depending on its features)
- Neural Network Classification ( label data depending on its features, i.e. tell whether or not a person has breast cancer )
- Computer Vision ( perform the same kind of tasks mentioned right above, but using images instead of tabular data )
- Transfer Learning ( when you want to solve a problem involving AI, it’s not impossible that someone has worked on a similar problem. Then, you can fine-tune their models to your needs instead of having to make a model from scratch )
- Natural Language Processing ( same thing but with text )
- Time Series ( estimate the value of something in the future )
Daniel would walk you through good practices and a complete workflow you can use when performing these tasks for each of these chapters.
Milestone projects
Projects are great: they are the proof you’re able to take a concept and apply it to a concrete problem. In this course, the milestone projects are more like walkthroughs rather than actual projects, but I’m not sure I would have been to finish these projects on my own. But fortunately, Daniel makes it look very easy and you don’t feel any pain doing these projects as he explains what you are coding does and how you could have found the solution by yourself.
Among these projects, you can find:
- Food101 ( replicate an actual paper to make a model able to recognize one dish among 101 recipes )
- Skimlit ( takes as an input the introduction of a scientific article and split it into different paragraphs such as background, and context…. so you can skim through it )
- BitPredict ( create a model that would theoretically predict Bitcoin’s value. Obviously, he mentions it’s a project for fun and it’s not that easy to predict prices, and that he’s not a financial advisor.
Appendix
These are bonus sections in case you lack proficiency with a specific library or concept used in the course, such as :
- Machine Learning
- Scikit-Learn
- Pandas
- Numpy
These are a bit less exciting (in my opinion) but they are considered fundamentals if you want to work in the Data Science field with Python. Fortunately, Daniel uses the same (effective) pedagogy as usual and gives you assignments so you can make sure you understood what you’ve been taught.
Conclusion
As an introductory course about Deep Learning and Tensorflow ( while preparing you for the Tensorflow Developer Certificate exam ), I would definitely recommend this course. The way of teaching is both dynamic and easy to follow, the content is delivered in a way any beginner can understand, and you’ll hardly find a course with so much content in one place.
In addition to all of that, this course allows you to get to know some good practices when it comes to coding, such as ways to properly prepare your data. These good practices will make a difference as a professional if you manage to integrate them into your code.
In my opinion, this course is 100% worth your money. Of course, you can check other reviews in order to make a solid decision, but at least I’d thank you for reading my review.
Good luck on your learning path :)
