Article 14664 of comp.ai.neural-nets: Xref: glinda.oz.cs.cmu.edu comp.ai.neural-nets:14664 Path: honeydew.srv.cs.cmu.edu!fs7.ece.cmu.edu!europa.eng.gtefsd.com!howland.reston.ans.net!xlink.net!fauern!news.th-darmstadt.de!hrz-ws11.hrz.uni-kassel.de!neuro.informatik.uni-kassel.de!jochenr From: jochenr@neuro.informatik.uni-kassel.de (Jochen Ruhland) Newsgroups: comp.ai.neural-nets Subject: Re: Info on NN - shareware software packages (wanted) Date: 7 Feb 1994 12:10:15 GMT Organization: FG Neuronale Netzwerke - Uni Kassel Lines: 40 Distribution: na Message-ID: <2j5b37$iaj@hrz-ws11.hrz.uni-kassel.de> References: <1994Feb4.234846.6427@sifon.cc.mcgill.ca> Reply-To: jochenr@neuro.informatik.uni-kassel.de NNTP-Posting-Host: nick-0.neuro.informatik.uni-kassel.de You may fetch our PC-based BP simulater, called BackProp 1.4 It's a mouse driven program which lets you design your nets in an intuitive, graphical way. This limits it's use to nets of a smaller size (<25), but it'll do for learning purpose. It's implemeting some different BP-rules, including CG. It works for any feedforward structured nets, which must not be ordered in layers. The program comes with no printed doc, but is equipped with a context sensitive help facility. A VGA-card is recommended, EGA will do also. You can find it on nick-2.neuro.informatik.uni-kassel.de is 141.51.188.3, Aliases: ftp.neuro.informatik.uni-kassel.de in /pub/NeuralNets/BackProp-1.4 where you find a LHA packed file. If you don't have LHA to unpack it, you'll find a LHA-package also on our server. The program runs on machines with and without a numerical coprozessor (x87). For now only a binary is avaible, maybe we'll distribute the NN-part of the prgram as PASCAL-source in the future. ** Since we have a large wold-map in our office where we collect ** ** all the sites where people use our BackProp, please send us a ** ** postcard so that we can put some more needles in the map. ** -- * FG Neuronale Netzwerke / Uni Kassel * * Jochen Ruhland * * Heinrich-Plett-Str. 40 * * D-34132 Kassel * * jochenr@neuro.informatik.uni-kassel.de * * Tel: +49-561-804-4376 FAX: -4244 * /// Watson's extension to murphy's law: There's always one more PTF /// The AIX Lemma to Watson's extension: You'll always need one more pre-req