Closed deepaksuresh closed 4 years ago
Isn't the derivative supposed to be output*(1-output) rather than what's defined here?
output*(1-output)
>>deriv(CrossEntropyLoss(), 0,0.1) 1.1111111111111112
Shouldn't the derivative in this case be 0.09?
0.09
Isn't the derivative supposed to be
output*(1-output)
rather than what's defined here?Shouldn't the derivative in this case be
0.09
?