Hello Fastai community,
I have a question about the way we calculate the accuracy of mnist model here:
corrects = (preds>0.0).float()==train_y
Why we assume that 3’s will return a value>0.0 while 7’s are less that zero?
and one more thing, according to @arora_aman in his reading group on YT, this value “0.0” is a mistake and it should be “0.5”