Synopsis
		sfonelayer < x_train_obj.rsf ytrain=y_train_obj.rsf xval=x_val_obj.rsf yval=y_val_obj.rsf > loss_train_obj.rsf lossval=loss_val_obj.rsf

Reproducing the training process of the following example:
https://github.com/seg/tutorials-2018/tree/master/1808_Neural_networks
This is a minimal one-layer network:
x -> a -> z == y
The training uses L2 loss function.

Parameters
		file lossval= auxiliary output file name   file xval= auxiliary input file name   file ytrain= auxiliary input file name   file yval= auxiliary input file name