Neuron outputting more than 1.0f
The neuron in my library is outputting more than 1.0f which i find very weird, so clearly im doing something wrong. Ignore that one of the inputs is 4 on the github, that was fixed same issue still occurs though. any help is appreciated
https://github.com/DancingRicardo/DynamicML
GitHub
GitHub - DancingRicardo/DynamicML
Contribute to DancingRicardo/DynamicML development by creating an account on GitHub.
7 Replies
making all the inputs 1 makes the effect even more noticeable
(you shouldn't commit bin and obj in the git repo)
you can easily debug this with the ide going one line at a time looking at what
summation is doingfinally got it
been a lil busy lol
thats the output of the sums
which is what i expect
the output is practically identical when i output the result of the activation functions so clearly theyre not working
idk why though this should work fine
never mind, sigmoid actually works, relu does not though