To train a network with the Rosenblatt or Rumelhart software, the training file must be in the form of what we call a 'netfile'. Some examples are provided below -- a netfile is just a text file that defines network structure, defines the input patterns, and defines the output patterns. A short chapter from a book about using the software is also provided to give a description of the structure of the netfile. The book is Dawson, M.R.W. (2005) Connectionism: A Hands-On Approach. Oxford, UK. Important note: In my netfiles, the first input unit (often the first pitch-class) is A!

	Short chapter on training set structure from Dawson (2005)
	ScaleTonics.pdf	Class exercise describing training a perceptron to identify scale tonics.
	ScaleTonics.net	Inputs: 24 different scales defined by 7 pitch-classes. Output: Tonic of the scale
	ScaleMode.net	Inputs: 24 different scales defined by 7 pitch-classes. Output: Is scale major or minor?
	ClassifyTetra.pdf	Class exercise describing training networks to classify four different types of tetrachords
	FourTetras.net	Netfile for trainng network to classify four different types of tetrachords
	ClassifyExtended.pdf	Class exercise describing training multilayer perceptrons on extended tetrachords
	ExtendedTetra.net	Input a tetrachord and classify it into one of 12 different types.
	ii-V-I Progression (TwoFiveOne.net)	Inputs: Chord in this progression. Outputs: The next chord.