Would consider this learning.
Well i was learning machine learning model from "hands on machine learning" book. I was doing all the implementation of linear regression , softmax regression from scratch, however when i entered the SVM chapter it really didn't talk much about the implementation or the maths behind it. Having taken advanced calculus and linear algebra in my first and fourth semester i thought the math wouldn't be hard so i started to read the "Mathematics for Machine Learning" book i went into the SVM chapter and read through the chapter honestly the math didn't scare me off and i implemented the loss function view of primal SVM, then when i had to implement the Dual Support Vector Machine i couldn't do it.
Googled a bit and stumbled across a method called SMO for quadratic programming problems. I read through this one paper from microsoft. Honestly i understood the steps and how to do it but i didn't for the love of god understand why it was done a certain way. I did implement it using the pseudo code they had lying around in that paper ,however i couldn't understand the reason behind those steps.
So what should i do about it. Should i go back and try to understand it. Is it bad that i was afraid of the complexity of the algorithm ?