Which Framework to choose?

I have just finished the my first course and read about all these Frameworks like tensorflow, pytorch, matlab, scikit learn … My questions are: Is it possible to go wrong when I just choose one and then follow the path for a while till I get confident in using it? Is there one which is best to start with? How hard or easy is it to switch frameworks after a while?

Hi @Pascal_Winterle , I would say that between Pytorch and Tensorflow, you could choose the one you feel more comfortable with, but the best option is to try to be updated with both because we use to find situations where we find one or another. Regarding scikit, yes it’s good to have experience as well because it can help you automate some tasks, and Matlab will facilitate some visualizations. So, you can see that, depending on your scenario, if you have a good knowledge of the tools, that will simplify your life a lot.

1 Like

Welcome to the community @Pascal_Winterle!

Great question!

I would argue that it’s more important to understand the underlying concepts and first principles of ML than to chose the best framework as of today. Also also believe that in 5-10 years the tool and technology landscape will change a lot. So what was the „Jupyter Nootbook yesterday“ might be „e.g. Databricks“ in an cloud environment today and maybe tomorrow might be another advanced framework covering also other aspects like federated learning or the focus on robust eco systems.

Still, here you can find a nice assessment from Sebastian Raschka on Linkedin with an overview on most recent frameworks in AI and how they correlate to winning e.g. Kaggle competitions:

Keep in mind that ecosystem and tool landscape keep evolving quite quickly in AI.

Happy learning!

Best regards

1 Like

There is no way to go wrong by choosing one framework and stick in it in order to get specialized. In fact, this can be a good approach. By focusing on one framework initially, you can build a solid foundation and gain confidence in using it effectively. It’s often better to become proficient in one framework rather than trying to learn them all at once. However, keep in mind that different frameworks have different strengths and use cases, so your choice should align with your goals and interests.

Keep in mind that these are just tools. Different ways to do the same task. I would focus in understand the foundation concepts and the mathematics behind the machine learning techniques instead.

Best regards