#====================================================

*call	nn	net

nodes	input	3		-0.5	0.5	in-layer
nodes	output	2		-0.5	0.5	out-layer

links	active	0	0	in-layer	out-layer

#	This call creates an ANN named "net".

#	The input layer is called "in-layer", the output layer is
#	"out-layer". Response functions are sigmoidal (default), the
#	starting biases are random values unifirmly distributed between
#	-0.5 and 0.5.

#	The list "links" connects these layers with active (adjustable during
#	training) weights.  The original values of the weights are all 0.