1. Training  //
==========
There are two steps in training a neural network:
1.1 Preparing the neural network inputs
	Following data is required in the WORKING directory.
	---------------------------------------------------
	(i)   nnbench.rc (Just the name of the directory where params are written) e.g. /home/shandar/instdir/params
	(ii)  list.codes (Names of the PDB codes). There is no restriction 
		on the way names are written but all feature data files will
		start with the names written here. Look at example/list.codes
	(iii) list.props (names of additional properties data, 
			if any e.g. asa, secondary structure)
		File format:
			Field 1: Type of the property, category or numerical
			Field 2: Number of properties of this type 
			Field 3-N : Names of the file extensions corresponding 
					to the properties mentioned there.
		Look at example/list.props
	(iv) pssm data. All PSSM data should be in pssm directory 
		as a subdiectory in the present working directory. 
		File names should be $code.pssm, where $code is the exact name
		used from the file list.codes.
		(Look at example directory)
	(v) props directory has all the data corresponding to list.props
		Data should have the same file names as $code in list.codes,
		with extensions reprsenting the property as listed in 
		list.props. 
		Inside each file, two column vertical format should be used.
		First column is the residue name and second column is the name
		of the category or numerical value of the property.
		Look at example/props directory.
	(vi) targets directory has the information on the target property to
		predicted. All files should be names as $code.o, and contain
		information in two column format. Only numerical values can
		be used in the current version.

	Following Script will prepare the neural network inputs data
	------------------------------------------------------------
	(i) get-20bits.sh will read the data from two column files 
	    e.g. targets/$code.o and prepare 20-bit codification files.
	(ii)	

1.2 Setting training and network architecture options.
1.3 Carrying out training.

 
