10 comments

[ 3.3 ms ] story [ 41.6 ms ] thread
"As Jeremy Howard points out, even academic papers often use softmax for multi-class classification, and I too have already seen it used incorrectly in blogs and papers during my short time studying DL."

AFAIK softmax should be used with mutli-class classification and sigmoid can be used with mutli-label classification.

Exactly. I’m pretty confused about the excerpt, it struck me as strange as soon as I read it - maybe it’s a typo?
I think the author meant sigmoid.
I think the author meant multi-label.
Nice, but doesn't do the Fast.ai course justice. All of the 10 things mentioned are shown in the first 3 / 14 lessons.
Fast.ai + Floydhub is something I’d be really interested in... but this appears to just be about the former?
Are you referring to using fast.ai library on FloydHub for you own project or going through the fast.ai course notebooks?
Is the page only blurry to me?
Were you on your phone when visiting the page?