Execution Time0.51s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-ubuntu18-gcc7 (sft-ubuntu-1804-3) on 2019-11-13 02:11:31
Repository revision: 30660dce2d9e89e4852dbf83dbd8b2cfcc137eff

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 var00.0469919
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1154      -1.496 
   1 |    -0.1429     -0.4034 
   2 |      0.504     -0.2182 
   3 |   -0.06663       2.191 
   4 |     0.1174     -0.7432 
   5 |     0.2033      -2.431 
   6 |    -0.7032     -0.2464 
   7 |     0.5065     -0.6636 
   8 |   0.001662      -3.516 
   9 |   -0.06555      0.5274 

output BN 
output DL feature 0 mean 0.0469919	output DL std 0.346483
output DL feature 1 mean -0.699948	output DL std 1.56191
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      0.208     -0.5372 
   1 |    -0.5775      0.2001 
   2 |       1.39      0.3251 
   3 |    -0.3455       1.951 
   4 |      0.214    -0.02917 
   5 |     0.4754      -1.168 
   6 |     -2.281      0.3061 
   7 |      1.397     0.02452 
   8 |    -0.1378      -1.901 
   9 |    -0.3422      0.8283 

output BN feature 0 mean 4.996e-17	output BN std 1.05361
output BN feature 1 mean 8.88178e-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 |     -0.454     -0.0917     -0.4738     -0.4161 
   1 |  -0.006666     0.04808     0.04695      -0.118 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    0.03223      0.1922      0.2015     -0.3069 
   1 |     -1.265     -0.5789     -0.6363     -0.4458 

 training batch 2 mu var00.0469948
compute loss for weight  0.0322417  0.0322317 result 1.10444
 training batch 3 mu var00.0469919
compute loss for weight  0.0322217  0.0322317 result 1.10445
 training batch 4 mu var00.0469927
compute loss for weight  0.0322367  0.0322317 result 1.10444
 training batch 5 mu var00.0469919
compute loss for weight  0.0322267  0.0322317 result 1.10445
   --dy = -0.454007 dy_ref = -0.454007
 training batch 6 mu var00.0469915
compute loss for weight  0.192203  0.192193 result 1.10445
 training batch 7 mu var00.0469919
compute loss for weight  0.192183  0.192193 result 1.10445
 training batch 8 mu var00.0469918
compute loss for weight  0.192198  0.192193 result 1.10445
 training batch 9 mu var00.0469919
compute loss for weight  0.192188  0.192193 result 1.10445
   --dy = -0.0917007 dy_ref = -0.0917007
 training batch 10 mu var00.0469923
compute loss for weight  0.20151  0.2015 result 1.10444
 training batch 11 mu var00.0469919
compute loss for weight  0.20149  0.2015 result 1.10445
 training batch 12 mu var00.0469921
compute loss for weight  0.201505  0.2015 result 1.10444
 training batch 13 mu var00.0469919
compute loss for weight  0.201495  0.2015 result 1.10445
   --dy = -0.473766 dy_ref = -0.473766
 training batch 14 mu var00.0469919
compute loss for weight  -0.306894  -0.306904 result 1.10444
 training batch 15 mu var00.0469919
compute loss for weight  -0.306914  -0.306904 result 1.10445
 training batch 16 mu var00.0469919
compute loss for weight  -0.306899  -0.306904 result 1.10444
 training batch 17 mu var00.0469919
compute loss for weight  -0.306909  -0.306904 result 1.10445
   --dy = -0.416117 dy_ref = -0.416117
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |   -0.01463       2.224 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.0469919
compute loss for weight  1.00001  1 result 1.10445
 training batch 19 mu var00.0469919
compute loss for weight  0.99999  1 result 1.10445
 training batch 20 mu var00.0469919
compute loss for weight  1.00001  1 result 1.10445
 training batch 21 mu var00.0469919
compute loss for weight  0.999995  1 result 1.10445
   --dy = -0.0146281 dy_ref = -0.0146281
 training batch 22 mu var00.0469919
compute loss for weight  1.00001  1 result 1.10447
 training batch 23 mu var00.0469919
compute loss for weight  0.99999  1 result 1.10442
 training batch 24 mu var00.0469919
compute loss for weight  1.00001  1 result 1.10446
 training batch 25 mu var00.0469919
compute loss for weight  0.999995  1 result 1.10444
   --dy = 2.22352 dy_ref = 2.22352
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  2.429e-17   1.943e-16 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.0469919
compute loss for weight  1e-05  0 result 1.10445
 training batch 27 mu var00.0469919
compute loss for weight  -1e-05  0 result 1.10445
 training batch 28 mu var00.0469919
compute loss for weight  5e-06  0 result 1.10445
 training batch 29 mu var00.0469919
compute loss for weight  -5e-06  0 result 1.10445
   --dy = 0 dy_ref = 2.42861e-17
 training batch 30 mu var00.0469919
compute loss for weight  1e-05  0 result 1.10445
 training batch 31 mu var00.0469919
compute loss for weight  -1e-05  0 result 1.10445
 training batch 32 mu var00.0469919
compute loss for weight  5e-06  0 result 1.10445
 training batch 33 mu var00.0469919
compute loss for weight  -5e-06  0 result 1.10445
   --dy = 0 dy_ref = 1.94289e-16
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1223      -2.089 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.1196      -1.065 

 training batch 34 mu var00.0469919
compute loss for weight  -0.119589  -0.119599 result 1.10445
 training batch 35 mu var00.0469919
compute loss for weight  -0.119609  -0.119599 result 1.10445
 training batch 36 mu var00.0469919
compute loss for weight  -0.119594  -0.119599 result 1.10445
 training batch 37 mu var00.0469919
compute loss for weight  -0.119604  -0.119599 result 1.10445
   --dy = 0.122309 dy_ref = 0.122309
 training batch 38 mu var00.0469919
compute loss for weight  -1.06461  -1.06462 result 1.10443
 training batch 39 mu var00.0469919
compute loss for weight  -1.06463  -1.06462 result 1.10447
 training batch 40 mu var00.0469919
compute loss for weight  -1.06462  -1.06462 result 1.10444
 training batch 41 mu var00.0469919
compute loss for weight  -1.06463  -1.06462 result 1.10446
   --dy = -2.08856 dy_ref = -2.08856
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m2.79614e-09[NON-XML-CHAR-0x1B][39m