Execution Time0.07s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-fedora28-gcc8 (sft-fedora-28-1.cern.ch) on 2019-11-13 01:11:43

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 var0-0.0820785
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    0.08272       1.151 
   1 |    -0.1935      0.3074 
   2 |    -0.6156      0.3031 
   3 |      -2.57    -0.07418 
   4 |    -0.5433      0.9657 
   5 |    -0.5387       2.282 
   6 |     0.7976      0.7459 
   7 |      1.355       -1.36 
   8 |      1.783       1.904 
   9 |    -0.3787     -0.1201 

output BN 
output DL feature 0 mean -0.0820785	output DL std 1.21759
output DL feature 1 mean 0.610542	output DL std 1.05295
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1427      0.5413 
   1 |   -0.09643     -0.3034 
   2 |    -0.4619     -0.3078 
   3 |     -2.154     -0.6854 
   4 |    -0.3993      0.3555 
   5 |    -0.3953       1.674 
   6 |     0.7616      0.1355 
   7 |      1.244      -1.973 
   8 |      1.615       1.295 
   9 |    -0.2568     -0.7314 

output BN feature 0 mean 8.88178e-17	output BN std 1.05405
output BN feature 1 mean -5.55112e-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 |    -0.7762      0.6816     -0.5976     -0.3038 
   1 |     -2.651       1.077      -2.494      -2.093 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.2317      0.8568      0.3137      0.7126 
   1 |     0.6982     -0.5244      0.3021      0.3432 

 training batch 2 mu var0-0.0820757
compute loss for weight  0.23166  0.23165 result 1.8959
 training batch 3 mu var0-0.0820785
compute loss for weight  0.23164  0.23165 result 1.89592
 training batch 4 mu var0-0.0820778
compute loss for weight  0.231655  0.23165 result 1.89591
 training batch 5 mu var0-0.0820785
compute loss for weight  0.231645  0.23165 result 1.89591
   --dy = -0.776222 dy_ref = -0.776222
 training batch 6 mu var0-0.082079
compute loss for weight  0.856812  0.856802 result 1.89592
 training batch 7 mu var0-0.0820785
compute loss for weight  0.856792  0.856802 result 1.8959
 training batch 8 mu var0-0.0820787
compute loss for weight  0.856807  0.856802 result 1.89591
 training batch 9 mu var0-0.0820785
compute loss for weight  0.856797  0.856802 result 1.89591
   --dy = 0.681596 dy_ref = 0.681596
 training batch 10 mu var0-0.0820782
compute loss for weight  0.313691  0.313681 result 1.8959
 training batch 11 mu var0-0.0820785
compute loss for weight  0.313671  0.313681 result 1.89592
 training batch 12 mu var0-0.0820784
compute loss for weight  0.313686  0.313681 result 1.89591
 training batch 13 mu var0-0.0820785
compute loss for weight  0.313676  0.313681 result 1.89591
   --dy = -0.597614 dy_ref = -0.597614
 training batch 14 mu var0-0.0820785
compute loss for weight  0.712563  0.712553 result 1.89591
 training batch 15 mu var0-0.0820785
compute loss for weight  0.712543  0.712553 result 1.89591
 training batch 16 mu var0-0.0820785
compute loss for weight  0.712558  0.712553 result 1.89591
 training batch 17 mu var0-0.0820785
compute loss for weight  0.712548  0.712553 result 1.89591
   --dy = -0.303793 dy_ref = -0.303793
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      3.362      0.4298 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.0820785
compute loss for weight  1.00001  1 result 1.89594
 training batch 19 mu var0-0.0820785
compute loss for weight  0.99999  1 result 1.89588
 training batch 20 mu var0-0.0820785
compute loss for weight  1.00001  1 result 1.89593
 training batch 21 mu var0-0.0820785
compute loss for weight  0.999995  1 result 1.89589
   --dy = 3.36205 dy_ref = 3.36205
 training batch 22 mu var0-0.0820785
compute loss for weight  1.00001  1 result 1.89591
 training batch 23 mu var0-0.0820785
compute loss for weight  0.99999  1 result 1.89591
 training batch 24 mu var0-0.0820785
compute loss for weight  1.00001  1 result 1.89591
 training batch 25 mu var0-0.0820785
compute loss for weight  0.999995  1 result 1.89591
   --dy = 0.429764 dy_ref = 0.429764
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  2.325e-16  -6.592e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.0820785
compute loss for weight  1e-05  0 result 1.89591
 training batch 27 mu var0-0.0820785
compute loss for weight  -1e-05  0 result 1.89591
 training batch 28 mu var0-0.0820785
compute loss for weight  5e-06  0 result 1.89591
 training batch 29 mu var0-0.0820785
compute loss for weight  -5e-06  0 result 1.89591
   --dy = -3.70074e-12 dy_ref = 2.32453e-16
 training batch 30 mu var0-0.0820785
compute loss for weight  1e-05  0 result 1.89591
 training batch 31 mu var0-0.0820785
compute loss for weight  -1e-05  0 result 1.89591
 training batch 32 mu var0-0.0820785
compute loss for weight  5e-06  0 result 1.89591
 training batch 33 mu var0-0.0820785
compute loss for weight  -5e-06  0 result 1.89591
   --dy = -1.85037e-11 dy_ref = -6.59195e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -2.549      0.8219 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.319      0.5229 

 training batch 34 mu var0-0.0820785
compute loss for weight  -1.31896  -1.31897 result 1.89588
 training batch 35 mu var0-0.0820785
compute loss for weight  -1.31898  -1.31897 result 1.89594
 training batch 36 mu var0-0.0820785
compute loss for weight  -1.31896  -1.31897 result 1.8959
 training batch 37 mu var0-0.0820785
compute loss for weight  -1.31897  -1.31897 result 1.89592
   --dy = -2.549 dy_ref = -2.549
 training batch 38 mu var0-0.0820785
compute loss for weight  0.522915  0.522905 result 1.89592
 training batch 39 mu var0-0.0820785
compute loss for weight  0.522895  0.522905 result 1.8959
 training batch 40 mu var0-0.0820785
compute loss for weight  0.52291  0.522905 result 1.89591
 training batch 41 mu var0-0.0820785
compute loss for weight  0.5229  0.522905 result 1.89591
   --dy = 0.821878 dy_ref = 0.821878
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m2.57048e-10[NON-XML-CHAR-0x1B][39m