feedforward-neural-networks

feedforward-neural-networks copied to clipboard

activation function option

2

currently we can only have the same activation function for all layers. is it normal for FNN or the option is just missing?

targos

There are problems when we have many output nodes. As you can see in the tests, one of them is skipped because it's not possible to find the good result...

jajoe

how to decide how many nodes are in the output layer?

1

from the tests, it seems that we have one output node

targos

is there a notion of convergence?

2

the iterations parameter litterally means that it will do X iterations. is there a way to know that the model has converged and stop before?

targos

danil-z