README.md
-
clone the repository: > git clone https://[email protected]/josefK128/backpropagation.git
-
install python 3.X (current latest 3.8) - see https://www.python.org/downloads/ - simply click download button and follow defaults for installation
-
install numpy: > pip install numpy
-
backprop222.py usage: > py backprop.py 1 True (>log) This command will write out all diagnostics to the console (or to a logfile)
-
backpropIJK.py usage: > py backpropIJK.py I J K 1 True (>log)
-
to run backpropIJK to verify correctness compare to out put of backprop222.py by running the following: > py backpropIJK.py 2 2 2 1 True
(Also see and compare the two log files: backprop222_diagnostics and backpropIJK-222_diagnostics)
-
For both backprop222.py and backpropIJK.py, to view convergence of the error function E to 0 run several sessions with increasing number of iterations (best to use diagnostics=False)