In order to do backpropagation and optimization, we need to have some measure of how wrong the model is. For this, we use a loss function. In our case, with a softmax classifier, we'll be using categorical crossentropy.

Next video:    • Neural Networks from Scratch  P.8 Im...  

Neural Networks from Scratch book: https://nnfs.io

Playlist for this series:    • Neural Networks from Scratch in Python  

Channel membership:    / @sentdex  
Discord:   / discord  
Support the content: https://pythonprogramming.net/support...
Twitter:   / sentdex  
Instagram:   / sentdex  
Facebook:   / pythonprogramming.net  
Twitch:   / sentdex  

#nnfs #python #neuralnetworks