Using python

Actually i have basic knowledge of python and can understand the optional labs but there are many course bases libraries used in the labs and I’m not able to get that and finding difficulties in plotting the data moreover according to you what is preferable among writing the complete code for regression or only learning about the libraries used for machine learning.

1 Like

Hi @Rohit_Kumar29, welcome to the community!

I struggled with the same challenge, coming from a different programming language.

For the short term, I focused on understanding just enough about the libraries to deliver the assignments.
Still, I kept meticulous notes about the things I didn’t understand well and then took the time to work out these aspects in the long run.

My idea was not to block or delay my learning but to take action to eventually create a deeper understanding.

You’re not alone. Ask the community if you’re too confused about anything.

1 Like

Dear @Rohit_Kumar29 ,

What problem exactly are you facing?

Over time you will want to improve your python skills. It is true that everyone uses libraries and frameworks to build real solutions these days: the state of the art is so complicated, that nobody has the time to build everything from scratch. That’s what the frameworks are for. But you still have to write code to implement things using TensorFlow or PyTorch, so whatever time you spend learning python will definitely not be wasted. Also note that the frameworks are very complex and take full advantage of the Object Oriented features of python, so you need more than just a basic knowledge of python.

The other point is that the way Prof Ng teaches us about ML/DL, he will always explain how the algorithm works and what is really doing. The lectures and the assignments will lead you through building the core algorithms “by hand” in python. The reason is that gives us a much better understanding and intuition when we switch to building systems using TensorFlow. If you just view all the algorithms as “black boxes” with mysterious stuff happening “under the covers”, then it’s harder to understand how to approach solving real problems. So the bottom line is that you need to be able to follow the things he is teaching us both in terms of concepts and in terms of how to code things. Then those coding skills will still be required even if you are using TF to build a solutions to complex problems.

1 Like

Hi @Rohit_Kumar29.

I can confirm, you are not alone. Couple of years ago I had exactly the same problem. After a while I decided to start with some Python courses. I can recommend you following course by Michigan University on Coursera.

You can use it for anything you want, not just the field of machine learning.


1 Like