Execution Time0.14s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-centos7-gcc62-opt-no-rt-cxxmodules (olhswep09.cern.ch) on 2019-11-14 13:04:25
Repository revision: 14de58de35eff907054671888ccc2de0f7f27e77

Test Timing: Passed
Processors1

Show Command Line
Display graphs:

Test output
Testing Backpropagation:
DEEP NEURAL NETWORK:   Depth = 3  Input = ( 1, 10, 4 )  Batch size = 10  Loss function = R
	Layer 0	 DENSE Layer: 	 ( Input =     4 , Width =     2 ) 	Output = (  1 ,    10 ,     2 ) 	 Activation Function = Identity
	Layer 1	 BATCH NORM Layer: 	 ( Input =     2 ) 
	Layer 2	 DENSE Layer: 	 ( Input =     2 , Width =     1 ) 	Output = (  1 ,    10 ,     1 ) 	 Activation Function = Identity
input 

10x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.9989     -0.4348      0.7818    -0.03005 
   1 |     0.8243    -0.05672     -0.9009     -0.0747 
   2 |   0.007912     -0.4108       1.391     -0.9851 
   3 |   -0.04894      -1.443      -1.061      -1.388 
   4 |     0.7674      -0.736      0.5797     -0.3821 
   5 |      2.061      -1.235       1.165     -0.4542 
   6 |    -0.1348     -0.4996     -0.1824       1.844 
   7 |    -0.2428       1.997    0.004806     -0.4222 
   8 |      1.541     0.09474       1.525       1.217 
   9 |    -0.1363     -0.1992     -0.2938     -0.1184 

 training batch 1 mu var00.978726
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      2.002       1.057 
   1 |      0.527        1.01 
   2 |     0.1313      -0.668 
   3 |     -1.991      -2.386 
   4 |      1.219      0.2112 
   5 |      3.484       1.588 
   6 |      1.286    -0.08262 
   7 |    -0.7583       1.765 
   8 |      4.393        3.03 
   9 |    -0.5052     -0.5098 

output BN 
output DL feature 0 mean 0.978726	output DL std 1.94639
output DL feature 1 mean 0.501404	output DL std 1.52348
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.5541      0.3841 
   1 |    -0.2446       0.352 
   2 |    -0.4589     -0.8091 
   3 |     -1.608      -1.998 
   4 |     0.1299     -0.2008 
   5 |      1.357      0.7519 
   6 |     0.1662     -0.4041 
   7 |    -0.9407      0.8742 
   8 |      1.849       1.749 
   9 |    -0.8036     -0.6996 

output BN feature 0 mean -8.88178e-17	output BN std 1.05408
output BN feature 1 mean -2.22045e-17	output BN std 1.05407
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    0.01414     -0.5196     0.01788    -0.04909 
   1 |    -0.6291      0.3035     -0.6219     -0.4689 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |      1.478    -0.01798       0.696       0.869 
   1 |      1.472       1.144      0.1216      0.3847 

 training batch 2 mu var00.978728
compute loss for weight  1.47768  1.47767 result 0.440729
 training batch 3 mu var00.978726
compute loss for weight  1.47766  1.47767 result 0.440729
 training batch 4 mu var00.978726
compute loss for weight  1.47767  1.47767 result 0.440729
 training batch 5 mu var00.978726
compute loss for weight  1.47766  1.47767 result 0.440729
   --dy = 0.0141409 dy_ref = 0.0141409
 training batch 6 mu var00.978725
compute loss for weight  -0.0179704  -0.0179804 result 0.440724
 training batch 7 mu var00.978726
compute loss for weight  -0.0179904  -0.0179804 result 0.440734
 training batch 8 mu var00.978725
compute loss for weight  -0.0179754  -0.0179804 result 0.440726
 training batch 9 mu var00.978726
compute loss for weight  -0.0179854  -0.0179804 result 0.440731
   --dy = -0.519573 dy_ref = -0.519573
 training batch 10 mu var00.978726
compute loss for weight  0.696051  0.696041 result 0.440729
 training batch 11 mu var00.978726
compute loss for weight  0.696031  0.696041 result 0.440729
 training batch 12 mu var00.978726
compute loss for weight  0.696046  0.696041 result 0.440729
 training batch 13 mu var00.978726
compute loss for weight  0.696036  0.696041 result 0.440729
   --dy = 0.0178765 dy_ref = 0.0178765
 training batch 14 mu var00.978726
compute loss for weight  0.869059  0.869049 result 0.440728
 training batch 15 mu var00.978726
compute loss for weight  0.869039  0.869049 result 0.440729
 training batch 16 mu var00.978726
compute loss for weight  0.869054  0.869049 result 0.440729
 training batch 17 mu var00.978726
compute loss for weight  0.869044  0.869049 result 0.440729
   --dy = -0.0490895 dy_ref = -0.0490895
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.6535       0.228 

weights for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |          1           1 

 training batch 18 mu var00.978726
compute loss for weight  1.00001  1 result 0.440735
 training batch 19 mu var00.978726
compute loss for weight  0.99999  1 result 0.440722
 training batch 20 mu var00.978726
compute loss for weight  1.00001  1 result 0.440732
 training batch 21 mu var00.978726
compute loss for weight  0.999995  1 result 0.440726
   --dy = 0.653503 dy_ref = 0.653503
 training batch 22 mu var00.978726
compute loss for weight  1.00001  1 result 0.440731
 training batch 23 mu var00.978726
compute loss for weight  0.99999  1 result 0.440727
 training batch 24 mu var00.978726
compute loss for weight  1.00001  1 result 0.44073
 training batch 25 mu var00.978726
compute loss for weight  0.999995  1 result 0.440728
   --dy = 0.227955 dy_ref = 0.227955
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -8.327e-17   7.633e-17 

weights for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |          0           0 

 training batch 26 mu var00.978726
compute loss for weight  1e-05  0 result 0.440729
 training batch 27 mu var00.978726
compute loss for weight  -1e-05  0 result 0.440729
 training batch 28 mu var00.978726
compute loss for weight  5e-06  0 result 0.440729
 training batch 29 mu var00.978726
compute loss for weight  -5e-06  0 result 0.440729
   --dy = -5.55112e-12 dy_ref = -8.32667e-17
 training batch 30 mu var00.978726
compute loss for weight  1e-05  0 result 0.440729
 training batch 31 mu var00.978726
compute loss for weight  -1e-05  0 result 0.440729
 training batch 32 mu var00.978726
compute loss for weight  5e-06  0 result 0.440729
 training batch 33 mu var00.978726
compute loss for weight  -5e-06  0 result 0.440729
   --dy = 0 dy_ref = 7.63278e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.6504     -0.2554 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.005     -0.8926 

 training batch 34 mu var00.978726
compute loss for weight  1.00473  1.00472 result 0.440735
 training batch 35 mu var00.978726
compute loss for weight  1.00471  1.00472 result 0.440722
 training batch 36 mu var00.978726
compute loss for weight  1.00473  1.00472 result 0.440732
 training batch 37 mu var00.978726
compute loss for weight  1.00472  1.00472 result 0.440726
   --dy = 0.650433 dy_ref = 0.650433
 training batch 38 mu var00.978726
compute loss for weight  -0.892569  -0.892579 result 0.440726
 training batch 39 mu var00.978726
compute loss for weight  -0.892589  -0.892579 result 0.440731
 training batch 40 mu var00.978726
compute loss for weight  -0.892574  -0.892579 result 0.440728
 training batch 41 mu var00.978726
compute loss for weight  -0.892584  -0.892579 result 0.44073
   --dy = -0.255389 dy_ref = -0.255389
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m9.60749e-10[NON-XML-CHAR-0x1B][39m