Execution Time0.06s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-fedora31-gcc9 (root-fedora-31-1.cern.ch) on 2019-11-13 00:48:28

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 var00.134116
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1657     -0.9869 
   1 |    -0.1059      0.4246 
   2 |      1.089      0.5494 
   3 |      1.282       1.601 
   4 |     0.5004     -0.5147 
   5 |     0.6677      -1.436 
   6 |     -1.424      -3.596 
   7 |    -0.0487       2.749 
   8 |    -0.8887      -3.154 
   9 |     0.1037      0.1973 

output BN 
output DL feature 0 mean 0.134116	output DL std 0.831023
output DL feature 1 mean -0.416591	output DL std 1.97448
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    0.04004     -0.3045 
   1 |    -0.3044      0.4491 
   2 |      1.211      0.5157 
   3 |      1.456       1.077 
   4 |     0.4645    -0.05236 
   5 |     0.6768      -0.544 
   6 |     -1.976      -1.697 
   7 |    -0.2319        1.69 
   8 |     -1.297      -1.462 
   9 |   -0.03853      0.3277 

output BN feature 0 mean 6.31439e-17	output BN std 1.05401
output BN feature 1 mean 0	output BN std 1.05408
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.9198      -1.383      0.9655      0.4478 
   1 |   -0.06186      0.4432    -0.07427      0.8037 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |   -0.06007      -0.205      0.1433     -0.8179 
   1 |    -0.1792      0.9862     -0.5522       -1.75 

 training batch 2 mu var00.134119
compute loss for weight  -0.0600596  -0.0600696 result 1.02577
 training batch 3 mu var00.134116
compute loss for weight  -0.0600796  -0.0600696 result 1.02575
 training batch 4 mu var00.134117
compute loss for weight  -0.0600646  -0.0600696 result 1.02576
 training batch 5 mu var00.134116
compute loss for weight  -0.0600746  -0.0600696 result 1.02575
   --dy = 0.919756 dy_ref = 0.919756
 training batch 6 mu var00.134116
compute loss for weight  -0.204949  -0.204959 result 1.02575
 training batch 7 mu var00.134116
compute loss for weight  -0.204969  -0.204959 result 1.02577
 training batch 8 mu var00.134116
compute loss for weight  -0.204954  -0.204959 result 1.02575
 training batch 9 mu var00.134116
compute loss for weight  -0.204964  -0.204959 result 1.02577
   --dy = -1.38312 dy_ref = -1.38312
 training batch 10 mu var00.134117
compute loss for weight  0.143275  0.143265 result 1.02577
 training batch 11 mu var00.134116
compute loss for weight  0.143255  0.143265 result 1.02575
 training batch 12 mu var00.134116
compute loss for weight  0.14327  0.143265 result 1.02576
 training batch 13 mu var00.134116
compute loss for weight  0.14326  0.143265 result 1.02575
   --dy = 0.965535 dy_ref = 0.965535
 training batch 14 mu var00.134116
compute loss for weight  -0.817877  -0.817887 result 1.02576
 training batch 15 mu var00.134116
compute loss for weight  -0.817897  -0.817887 result 1.02575
 training batch 16 mu var00.134116
compute loss for weight  -0.817882  -0.817887 result 1.02576
 training batch 17 mu var00.134116
compute loss for weight  -0.817892  -0.817887 result 1.02576
   --dy = 0.447753 dy_ref = 0.447753
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |       2.18     -0.1285 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.134116
compute loss for weight  1.00001  1 result 1.02578
 training batch 19 mu var00.134116
compute loss for weight  0.99999  1 result 1.02574
 training batch 20 mu var00.134116
compute loss for weight  1.00001  1 result 1.02577
 training batch 21 mu var00.134116
compute loss for weight  0.999995  1 result 1.02575
   --dy = 2.18004 dy_ref = 2.18004
 training batch 22 mu var00.134116
compute loss for weight  1.00001  1 result 1.02576
 training batch 23 mu var00.134116
compute loss for weight  0.99999  1 result 1.02576
 training batch 24 mu var00.134116
compute loss for weight  1.00001  1 result 1.02576
 training batch 25 mu var00.134116
compute loss for weight  0.999995  1 result 1.02576
   --dy = -0.128517 dy_ref = -0.128517
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  4.025e-16  -9.021e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.134116
compute loss for weight  1e-05  0 result 1.02576
 training batch 27 mu var00.134116
compute loss for weight  -1e-05  0 result 1.02576
 training batch 28 mu var00.134116
compute loss for weight  5e-06  0 result 1.02576
 training batch 29 mu var00.134116
compute loss for weight  -5e-06  0 result 1.02576
   --dy = 3.33067e-11 dy_ref = 4.02456e-16
 training batch 30 mu var00.134116
compute loss for weight  1e-05  0 result 1.02576
 training batch 31 mu var00.134116
compute loss for weight  -1e-05  0 result 1.02576
 training batch 32 mu var00.134116
compute loss for weight  5e-06  0 result 1.02576
 training batch 33 mu var00.134116
compute loss for weight  -5e-06  0 result 1.02576
   --dy = -3.70074e-12 dy_ref = -9.02056e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.641     -0.1646 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.328      0.7807 

 training batch 34 mu var00.134116
compute loss for weight  -1.32814  -1.32815 result 1.02574
 training batch 35 mu var00.134116
compute loss for weight  -1.32816  -1.32815 result 1.02578
 training batch 36 mu var00.134116
compute loss for weight  -1.32814  -1.32815 result 1.02575
 training batch 37 mu var00.134116
compute loss for weight  -1.32815  -1.32815 result 1.02577
   --dy = -1.64141 dy_ref = -1.64141
 training batch 38 mu var00.134116
compute loss for weight  0.780668  0.780658 result 1.02576
 training batch 39 mu var00.134116
compute loss for weight  0.780648  0.780658 result 1.02576
 training batch 40 mu var00.134116
compute loss for weight  0.780663  0.780658 result 1.02576
 training batch 41 mu var00.134116
compute loss for weight  0.780653  0.780658 result 1.02576
   --dy = -0.164626 dy_ref = -0.164626
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m2.63637e-10[NON-XML-CHAR-0x1B][39m