Execution Time0.05s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-ubuntu14-gcc48 (ec-ubuntu-14-04-x86-64-2) on 2019-11-13 00:57:18
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 var0-0.454941
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.8595     -0.1565 
   1 |    -0.1275    -0.08705 
   2 |    -0.8584      0.9623 
   3 |   -0.09995      0.5111 
   4 |    -0.7895     0.02505 
   5 |     -1.749     -0.1565 
   6 |     0.6862      -2.234 
   7 |     0.1178       1.586 
   8 |     -1.003      -1.092 
   9 |     0.1331    -0.01932 

output BN 
output DL feature 0 mean -0.454941	output DL std 0.716665
output DL feature 1 mean -0.0661026	output DL std 1.04889
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.5949    -0.09081 
   1 |     0.4816    -0.02105 
   2 |    -0.5934       1.033 
   3 |     0.5221      0.5801 
   4 |     -0.492      0.0916 
   5 |     -1.903    -0.09089 
   6 |      1.678      -2.179 
   7 |     0.8423       1.661 
   8 |    -0.8055      -1.031 
   9 |     0.8648     0.04701 

output BN feature 0 mean -2.22045e-17	output BN std 1.05398
output BN feature 1 mean 2.01228e-17	output BN std 1.05404
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |      1.101      -1.354      0.5533       2.798 
   1 |      1.869      -1.162       1.752      0.2893 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    -0.5312     0.06143     -0.3744      0.3129 
   1 |   -0.03283      0.5685      0.1177      -1.048 

 training batch 2 mu var0-0.454938
compute loss for weight  -0.531236  -0.531246 result 4.88154
 training batch 3 mu var0-0.454941
compute loss for weight  -0.531256  -0.531246 result 4.88152
 training batch 4 mu var0-0.45494
compute loss for weight  -0.531241  -0.531246 result 4.88154
 training batch 5 mu var0-0.454941
compute loss for weight  -0.531251  -0.531246 result 4.88153
   --dy = 1.10099 dy_ref = 1.10099
 training batch 6 mu var0-0.454941
compute loss for weight  0.061436  0.061426 result 4.88152
 training batch 7 mu var0-0.454941
compute loss for weight  0.061416  0.061426 result 4.88155
 training batch 8 mu var0-0.454941
compute loss for weight  0.061431  0.061426 result 4.88153
 training batch 9 mu var0-0.454941
compute loss for weight  0.061421  0.061426 result 4.88154
   --dy = -1.35405 dy_ref = -1.35405
 training batch 10 mu var0-0.45494
compute loss for weight  -0.37435  -0.37436 result 4.88154
 training batch 11 mu var0-0.454941
compute loss for weight  -0.37437  -0.37436 result 4.88153
 training batch 12 mu var0-0.454941
compute loss for weight  -0.374355  -0.37436 result 4.88153
 training batch 13 mu var0-0.454941
compute loss for weight  -0.374365  -0.37436 result 4.88153
   --dy = 0.553309 dy_ref = 0.553309
 training batch 14 mu var0-0.454941
compute loss for weight  0.312882  0.312872 result 4.88156
 training batch 15 mu var0-0.454941
compute loss for weight  0.312862  0.312872 result 4.8815
 training batch 16 mu var0-0.454941
compute loss for weight  0.312877  0.312872 result 4.88155
 training batch 17 mu var0-0.454941
compute loss for weight  0.312867  0.312872 result 4.88152
   --dy = 2.79816 dy_ref = 2.79816
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |        1.2       8.563 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.454941
compute loss for weight  1.00001  1 result 4.88154
 training batch 19 mu var0-0.454941
compute loss for weight  0.99999  1 result 4.88152
 training batch 20 mu var0-0.454941
compute loss for weight  1.00001  1 result 4.88154
 training batch 21 mu var0-0.454941
compute loss for weight  0.999995  1 result 4.88153
   --dy = 1.19989 dy_ref = 1.19989
 training batch 22 mu var0-0.454941
compute loss for weight  1.00001  1 result 4.88162
 training batch 23 mu var0-0.454941
compute loss for weight  0.99999  1 result 4.88145
 training batch 24 mu var0-0.454941
compute loss for weight  1.00001  1 result 4.88157
 training batch 25 mu var0-0.454941
compute loss for weight  0.999995  1 result 4.88149
   --dy = 8.56317 dy_ref = 8.56317
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -7.633e-17   9.437e-16 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.454941
compute loss for weight  1e-05  0 result 4.88153
 training batch 27 mu var0-0.454941
compute loss for weight  -1e-05  0 result 4.88153
 training batch 28 mu var0-0.454941
compute loss for weight  5e-06  0 result 4.88153
 training batch 29 mu var0-0.454941
compute loss for weight  -5e-06  0 result 4.88153
   --dy = -1.33227e-10 dy_ref = -7.63278e-17
 training batch 30 mu var0-0.454941
compute loss for weight  1e-05  0 result 4.88153
 training batch 31 mu var0-0.454941
compute loss for weight  -1e-05  0 result 4.88153
 training batch 32 mu var0-0.454941
compute loss for weight  5e-06  0 result 4.88153
 training batch 33 mu var0-0.454941
compute loss for weight  -5e-06  0 result 4.88153
   --dy = 2.36848e-10 dy_ref = 9.4369e-16
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.888      -4.236 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.6357      -2.021 

 training batch 34 mu var0-0.454941
compute loss for weight  0.635703  0.635693 result 4.88155
 training batch 35 mu var0-0.454941
compute loss for weight  0.635683  0.635693 result 4.88151
 training batch 36 mu var0-0.454941
compute loss for weight  0.635698  0.635693 result 4.88154
 training batch 37 mu var0-0.454941
compute loss for weight  0.635688  0.635693 result 4.88152
   --dy = 1.88753 dy_ref = 1.88753
 training batch 38 mu var0-0.454941
compute loss for weight  -2.02139  -2.0214 result 4.88149
 training batch 39 mu var0-0.454941
compute loss for weight  -2.02141  -2.0214 result 4.88157
 training batch 40 mu var0-0.454941
compute loss for weight  -2.0214  -2.0214 result 4.88151
 training batch 41 mu var0-0.454941
compute loss for weight  -2.02141  -2.0214 result 4.88155
   --dy = -4.23625 dy_ref = -4.23625
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m2.36847e-10[NON-XML-CHAR-0x1B][39m