next up previous contents
Next: The Network Description File Up: Example Files Previous: Example Files

A Typical Logfile

  The class CMul2L writes a very detailed logfile. This is very useful for evaluation of experiments. All relevant information can be found in this file. The used text format is easy to read for the human user. It can also be used with text filters like grep or awk to extract information from a number of files.

 This is the logfile of the

 M O D U L A R  N E U R A L  N E T W O R K

                                   program


 Running Version: 1.0

Initialized at Thu Aug 22 13:06:04 1996
with the following pararmeters:
# of user inputs       : 80
# of different classes : 2

# of nets in the inp layer     : 5
# inputs in inp layer          : 16
# layers in inp layer          : 2
# lambda in inp layer          : 1.000000
# function in inp layer        : 1
# neurons in 0. layer          : 4
# outputs per net in inp layer : 1

# inputs in the decision net   : 5
# inputs in the decision net   : 0
# layers in the decision net   : 2
# lambda in the decision net   : 1.000000
# function in the decision net : 1
# neurons in 0. layer          : 4
# outputs in the decision net  : 2

this logfile name: multi.log
the mapping code 17
small intermediate representation

weights are initilalized randomly!


Training the whole Network !

Training the input layer !

Parameters: max. # of steps: 250
            max. Error     : 0.010000
            data file name : diab.t
            # train. recs. : 384
            error file name: input.err
            learning const.: 1.000000
            momentum const.: 0.000000

Start training at Thu Aug 22 13:06:39 1996

Stop training (input layer) at Thu Aug 22 13:39:57 1996

 Results : # of steps: 250
           Error     : 0.184489

Training the Decision Network !

Parameters: max. # of steps: 200
            max. Error     : 0.010000
            data file name : diab.t
            # train. recs. : 384
            error file name: decision.err
            learning const.: 1.000000
            momentum const.: 0.000000

Start training at Thu Aug 22 13:40:02 1996

Stop training (decision net) at Thu Aug 22 13:50:01 1996

Results : # of steps: 200
          Error     : 0.238476

Write the Network to file: diab80 in dir: nets
                  at Thu Aug 22 14:06:59 1996

Test the performance !

Parameters: data file name : diab.t
            # records      : 384
            difference     : 0.010000
            display mode   : 0

Start testing at Thu Aug 22 14:07:27 1996

 Test Statistics:

 Results: # records   : 384
          # goods     : 328
          # bads      : 56

          recognition : 85.416667 %

Test the performance !!!

Parameters: data file name : diab.p
            # records      : 384
            difference     : 0.010000
            display mode   : 0

Start testing at Thu Aug 22 14:07:48 1996

 Test Statistics:

 Results: # records   : 384
          # goods     : 288
          # bads      : 96

          recognition : 75.000000 %


The work is done ?! at Thu Aug 22 14:07:57 1996

 B Y E !!!


next up previous contents
Next: The Network Description File Up: Example Files Previous: Example Files

Albrecht Schmidt
Mit Okt 4 16:45:34 CEST 2000