function w = mlppak(net)%MLPPAK	Combines weights and biases into one weights vector.%%	Description%	W = MLPPAK(NET) takes a network data structure NET and combines the%	component weight matrices bias vectors into a single row vector W.%	The facility to switch between these two representations for the%	network parameters is useful, for example, in training a network by%	error function minimization, since a single vector of parameters can%	be handled by general-purpose optimization routines.%%	The ordering of the paramters in W is defined by%	  w = [net.w1(:)', net.b1, net.w2(:)', net.b2];%	 where W1 is the first-layer weight matrix, B1 is the first-layer%	bias vector, W2 is the second-layer weight matrix, and B2 is the%	second-layer bias vector.%%	See also%	MLP, MLPUNPAK, MLPFWD, MLPERR, MLPBKP, MLPGRAD%%	Copyright (c) Ian T Nabney (1996-2001)% Check arguments for consistencyerrstring = consist(net, 'mlp');if ~isempty(errstring);  error(errstring);endw = [net.w1(:)', net.b1, net.w2(:)', net.b2];