The backpropagation part is needed?
xuqingyu26 opened this issue · 1 comments
xuqingyu26 commented
I think with the definition of the network and loss function, the backpropagation should be auto computed, why should we explicitly definite it in the code?
RNGrunshen commented
I think with the definition of the network and loss function, the backpropagation should be auto computed, why should we explicitly definite it in the code?
because we are using numpy instead of torch, it is necessary for us to build the calculation diagram manually