Implementing Python Codes

Hi, everyone.

I recently started this course. I am in W2, C1. So far, it goes well - I understand the topics and the mathematics. I also understand the flow of most Python codes. The thing is if I am asked to implement those functions myself, it would be difficult for me. Especially, the parts where partial derivatives are converted to differences in Python, and when also you take on multiple variables level…

Anyways, is it normal to feel confused or there is something wrong goin on? I would appreciate if you share your opinions.




Hi @Mammad_Mammadov
Welcome to the community!

First It’s normal in the course to feel confused because you didn’t practise and doing many project with varios data but if you continue with the specialization this feel disappears will be that’s because you would learn many techniqes that deal with data and you would know the difference between when to use it besides doing many tasks in this course in adddition if you want to emphasize what you have learned and gain experience from other people you can create projects and read more notebooks(project) that done by other people on any platform like kaggle.Also Many of the problems and models now is built in in libraries so the implementation of it would be easy and if you want to implement it from scratch tasks in this specialization provide that

Best Regards,


Hello @Mammad_Mammadov,

Try to write the code yourself, making and modification and verification what happens this will help you understand the code better and write new code more easily.

I hope I could help you.

1 Like

Yeah, Kaggle is definitely a good idea. Thank you so much!

I will try to play with the parameters. Thanks a lot for the idea, appreciate it!

Hi @Mammad_Mammadov based on my experience I think is normal and very common, I got the same feeling when I was starting, but as everything is a skill that you will master with time and repetition. Often you will find yourself solving a problem that requires a function from the course you took and you will copy those functions, adapt to your use case and use it, as you needed more and more you will use it without realizing.

The feeling of not knowing is the first feeling of learning, so it might be overwhelming at first but as you practice everything will come into place and you will be able to implement stuff on your own.

I hope this helps

1 Like

Hi again, @pastorsoto .

Thank you so much for sharing your experience with me! It helps a lot to me to think of this as normal and believe in excelling on implementation over time.

1 Like

I agree with @pastorsoto . Also, Don’t be overwhelmed with the AI implementation geeks you encounter in github and youtube :slight_smile: We can certainly reach that level with rigorous practical implementation.
I made sure I implement everything in Python (I didnt even know python basics at that point) from ‘day 1’ of ML specialization…& I am 100% sure I took more time than my parallels who enroled in the same time.
But It was an effort which payed me off very well. I am starting to see the benfits now. But having said that…there were many instances in this learning & implementation timeline where I was stuckup umpteen number of times while implementing … a direct outcome of Cost-Benefit scenario!
Hope this helps in your decision making and cheers for great AI implementation ahead!


Hi again, @tennis_geek

Thanks a lot for sharing the experience and drawing the attention to the importance of rigorous practice. It definitely helps a lot!

Good luck to both of us on the way towards mastering ML!

So far I am on Week 1 and excuting Python program is ok. I have understanding of programming, and know basic Python.

For getting certificate, I see there are Labs. Would we have to write Python code ourselves?


You have to write your code in specific places only, where it is mentioned something like that:



You just need to write your code between these two lines and no need to write anywhere else or change any other code.


Yes, I see that. However, I kinda got curious why we were not required to write ourselves the other codes - the ones whose codes were already provided.


I think for the ones where we are explicitly asked to implement something, we must write the codes. However, for the ones that the instructors states there’s no need, we don’t have to. However, I am still curious why we don’t have to implement them. After all, we learn ML and I think we should know them as well.

Thank you Saif .
What would be the complexity level? Is there a time limit for those exams?

I may agree, but for Python beginners , this would be challenge to put entire code together. A simple video with all the ML packages would be useful.

Hi @Venkat_Subramani,

There is no time limit, and you might submit assignments as many times as you want. There are instructions that explain what are expected for each exercise, and beneath each exercise, there are also hints. If you are on the 7-day trial of this specialization, you might actually open one of the labs to see and try it for yourself, because it is the best way to know whether it is too complex for you.


Hello @Mammad_Mammadov,

After you completed the planned assignment, the rest is up to you. You might make a copy of the assignment, and then on the copy, choose some code cells that you want to challenge yourself, write down some key notes of what they should achieve, then remove those cells, and finally re-implement them on your own.

In this way you will not just experience more implementation work, but will also be required to understand how existing code works, and write down the key requirement that is sufficient for you to re-implement. You might also need to visit some online documentation to review additional packages used in the existing code that are not used in the exercises. Certainly this will take more time than the specialization has anticipated a learner needs to complete it.

Good luck!

1 Like

Thank you rmwkowk! I joined CourseEra and hoping to finish the course.
I love the teaching cadence and technique of Andrew Ng. Very well presented course.

Pls convey my thanks to him.


Good luck to you too, @Venkat_Subramani!