Jeremy mentioned in one of the lectures it’s good to learn Greek letter names, because it helps reading the math equations if we can verbalize them in our head. I think the same applies to reading fastai code. If I can deduce the variable/attribute/class name etc., it’s much easier to follow along. E.g. ftrs = features. I have a problem with many abbreviations though - I can’t make them into an English word. Could we do a Wiki post and crowdsource fastai code glossary?
Very cool, thanks for pointing! What would be the best way to add new abbreviations (to-be) defined? I’m reading mixup callback now, some examples:
lf - loss function?
lam - lambda?