OpenANN  1.1.0
An open source library for artificial neural networks.
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros Pages
Two Spirals

This benchmark is based on the example program Two Spirals.

The result will look like this:

$ ./TwoSpiralsBenchmark
Architecture: 2-20-10-1 (bias)
281 parameters
....................................................................................................
Finished 100 runs.
    Correct   Accuracy  Time/ms   Iterations
Mean+-StdDev  188.250+-1.554  0.975+-0.112  5981    768+-21.979
[min,max] [177,193] [0.917,1.000]     [206,4342]

Architecture: 2-20-20-1 (bias)
501 parameters
....................................................................................................
Finished 100 runs.
    Correct   Accuracy  Time/ms   Iterations
Mean+-StdDev  188.570+-1.448  0.977+-0.104  12466   464+-13.867
[min,max] [180,193] [0.933,1.000]     [192,1216]

Architecture: 2-20-20-1 (bias), Compression: 3-21-21
501 parameters
....................................................................................................
Finished 100 runs.
    Correct   Accuracy  Time/ms   Iterations
Mean+-StdDev  186.060+-1.533  0.964+-0.110  8174    305+-9.903
[min,max] [175,193] [0.907,1.000]     [153,1038]

Architecture: 2-20-20-1 (bias), Compression: 3-12-12
312 parameters
....................................................................................................
Finished 100 runs.
    Correct   Accuracy  Time/ms   Iterations
Mean+-StdDev  185.660+-1.709  0.962+-0.123  5248    511+-16.075
[min,max] [168,192] [0.870,0.995]     [192,2886]

Architecture: 2-20-20-1 (bias), Compression: 3-6-6
186 parameters
....................................................................................................
Finished 100 runs.
    Correct   Accuracy  Time/ms   Iterations
Mean+-StdDev  184.750+-1.914  0.957+-0.138  3033    679+-18.572
[min,max] [164,193] [0.850,1.000]     [209,3023]

Architecture: 2-20-20-1 (bias), Compression: 3-6-3
183 parameters
....................................................................................................
Finished 100 runs.
    Correct   Accuracy  Time/ms   Iterations
Mean+-StdDev  185.140+-1.798  0.959+-0.129  3381    775+-20.821
[min,max] [172,193] [0.891,1.000]     [234,6584]