weights-dec31-16_5.dat

足球机器人仿真组CMU97的源码

{{0.110625 -0.016381 -0.130543 0.140730 } {10.638384 -8.798143 6.113193 -7.543053 } {-0.174058 -1.035419 -0.303957 -0.944745 } {-4.862822 4.793542 -2.470693 4.103837 } }
{{19.317274 } {-16.828848 } {9.926309 } {-14.169587 } {-2.882147 } }

Last weights:
{{0.110625 -0.016381 -0.130543 0.140730 } {10.638384 -8.798143 6.113193 -7.543053 } {-0.174058 -1.035419 -0.303957 -0.944745 } {-4.862822 4.793542 -2.470693 4.103837 } }
{{19.317274 } {-16.828848 } {9.926309 } {-14.169587 } {-2.882147 } }

Last changes in weights:
{{-0.000006 -0.000009 0.000002 -0.000003 } {0.000060 -0.000035 0.000012 -0.000015 } {-0.000002 -0.000013 -0.000014 -0.000017 } {-0.000015 0.000036 0.000017 0.000034 } }
{{0.000009 } {-0.000006 } {-0.000013 } {0.000001 } {-0.000004 } }

The last forward pass, followed by target (epoch 3000):
0.985714 0.500000 0.200000 1.000000 
0.530520 0.171176 0.396214 0.281880 
0.629604 0.542690 0.597778 0.570007 1.000000 
-1.995703 
-1.995703 
18.040001 

dhid and dout from last pass:
128.960480 -117.397270 67.707626 -97.972816 
20.035704 

3->4->1 network:
	eta = (0.000001,0.000001), alpha = 0.900000, gamma = (0.999000,1.000000), randmax = 0.500000
	1462 training patterns from Data/Random.dat
Last Epoch 3000:  tot err = 404402.3750, avg err = 276.6090, num wrong = 1440
Best Epoch 3000:  tot err = 404402.3750, avg err = 276.6090, num wrong = 1440