To start with the ML or DL course


#1

As someone who only started out learning in this field a couple of months, and with a basic knowledge of ML and DL, I’ve been highly recommended to do the Fastai DL course. However, I now see that the ML course has been released and was wondering if that is something that should be done prior to the DL one? Has anybody been in a similar situation or have some advice they could share?


(Peter Marelas) #2

I’ve done both and I would recommend you start with ML. The first method discussed (Random Forests) is easier to grasp albeit very different to DL methods.

Peter


#3

You could do both in parallel if you have time, they both talk about similar things and from slightly different perspective.
BTW you can join online study group at TWiMLAI: https://twimlai.com/twiml-x-fast-ai/


(SA) #4

I had the same situation. Though, i don’t have much basic knowledge. I am starting with ML. Also, if I get time I will do the Live DL1 or see the archived recordings.
My only doubt is that DL1 2017 version is outdated (since fastai library v1 has been released) but what about the ML course? Is that dependent on old library?


(Marius) #5

actually I don’t find the fast.ai library that outdated for the DL1 course…I read the docs of the newest release of the fast.ai library and you understand it completly when learning it from the DL1 course. The function names might be a little be different, but the ideas are exactly the same!