Execution Time1.20s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-ubuntu18-clang91-dbg (sft-ubuntu-1804-3) on 2019-11-15 09:27:19
Repository revision: ddaf537cc1431ddcb8dc5b394576c131f26a6e1a

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 var00.327464
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.6723     -0.2962 
   1 |    -0.2895     -0.2469 
   2 |     0.6357      0.6835 
   3 |   -0.05157      0.6463 
   4 |     0.6317     -0.0542 
   5 |      1.275     -0.4285 
   6 |     0.4817      -1.331 
   7 |     -1.055      0.7198 
   8 |       1.05      -1.169 
   9 |   -0.07519     0.06702 

output BN 
output DL feature 0 mean 0.327464	output DL std 0.695043
output DL feature 1 mean -0.140927	output DL std 0.721306
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.5229     -0.2268 
   1 |    -0.9356     -0.1548 
   2 |     0.4674       1.205 
   3 |    -0.5748        1.15 
   4 |     0.4613      0.1267 
   5 |      1.436     -0.4202 
   6 |     0.2339      -1.739 
   7 |     -2.097       1.258 
   8 |      1.096      -1.502 
   9 |    -0.6106      0.3039 

output BN feature 0 mean -4.44089e-17	output BN std 1.05397
output BN feature 1 mean -1.11022e-17	output BN std 1.05398
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |   -0.01259    -0.03034     0.02629      -0.133 
   1 |    0.03742    -0.07897     0.08012    -0.09717 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.1216     -0.4755      0.4472      0.1855 
   1 |    -0.2849      0.1802     0.05902     -0.6879 

 training batch 2 mu var00.327467
compute loss for weight  0.121652  0.121642 result 0.321858
 training batch 3 mu var00.327464
compute loss for weight  0.121632  0.121642 result 0.321858
 training batch 4 mu var00.327465
compute loss for weight  0.121647  0.121642 result 0.321858
 training batch 5 mu var00.327464
compute loss for weight  0.121637  0.121642 result 0.321858
   --dy = -0.0125893 dy_ref = -0.0125893
 training batch 6 mu var00.327463
compute loss for weight  -0.475474  -0.475484 result 0.321857
 training batch 7 mu var00.327464
compute loss for weight  -0.475494  -0.475484 result 0.321858
 training batch 8 mu var00.327464
compute loss for weight  -0.475479  -0.475484 result 0.321858
 training batch 9 mu var00.327464
compute loss for weight  -0.475489  -0.475484 result 0.321858
   --dy = -0.0303432 dy_ref = -0.0303432
 training batch 10 mu var00.327464
compute loss for weight  0.447234  0.447224 result 0.321858
 training batch 11 mu var00.327464
compute loss for weight  0.447214  0.447224 result 0.321858
 training batch 12 mu var00.327464
compute loss for weight  0.447229  0.447224 result 0.321858
 training batch 13 mu var00.327464
compute loss for weight  0.447219  0.447224 result 0.321858
   --dy = 0.0262913 dy_ref = 0.0262913
 training batch 14 mu var00.327464
compute loss for weight  0.185521  0.185511 result 0.321856
 training batch 15 mu var00.327464
compute loss for weight  0.185501  0.185511 result 0.321859
 training batch 16 mu var00.327464
compute loss for weight  0.185516  0.185511 result 0.321857
 training batch 17 mu var00.327464
compute loss for weight  0.185506  0.185511 result 0.321858
   --dy = -0.132959 dy_ref = -0.132959
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |   -0.04831       0.692 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.327464
compute loss for weight  1.00001  1 result 0.321857
 training batch 19 mu var00.327464
compute loss for weight  0.99999  1 result 0.321858
 training batch 20 mu var00.327464
compute loss for weight  1.00001  1 result 0.321858
 training batch 21 mu var00.327464
compute loss for weight  0.999995  1 result 0.321858
   --dy = -0.0483052 dy_ref = -0.0483052
 training batch 22 mu var00.327464
compute loss for weight  1.00001  1 result 0.321865
 training batch 23 mu var00.327464
compute loss for weight  0.99999  1 result 0.321851
 training batch 24 mu var00.327464
compute loss for weight  1.00001  1 result 0.321861
 training batch 25 mu var00.327464
compute loss for weight  0.999995  1 result 0.321854
   --dy = 0.692021 dy_ref = 0.692021
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  1.735e-18  -2.429e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.327464
compute loss for weight  1e-05  0 result 0.321858
 training batch 27 mu var00.327464
compute loss for weight  -1e-05  0 result 0.321858
 training batch 28 mu var00.327464
compute loss for weight  5e-06  0 result 0.321858
 training batch 29 mu var00.327464
compute loss for weight  -5e-06  0 result 0.321858
   --dy = 2.77556e-12 dy_ref = 1.73472e-18
 training batch 30 mu var00.327464
compute loss for weight  1e-05  0 result 0.321858
 training batch 31 mu var00.327464
compute loss for weight  -1e-05  0 result 0.321858
 training batch 32 mu var00.327464
compute loss for weight  5e-06  0 result 0.321858
 training batch 33 mu var00.327464
compute loss for weight  -5e-06  0 result 0.321858
   --dy = 0 dy_ref = -2.42861e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.4738      -1.122 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.1019      -0.617 

 training batch 34 mu var00.327464
compute loss for weight  -0.101937  -0.101947 result 0.321863
 training batch 35 mu var00.327464
compute loss for weight  -0.101957  -0.101947 result 0.321853
 training batch 36 mu var00.327464
compute loss for weight  -0.101942  -0.101947 result 0.32186
 training batch 37 mu var00.327464
compute loss for weight  -0.101952  -0.101947 result 0.321855
   --dy = 0.473828 dy_ref = 0.473828
 training batch 38 mu var00.327464
compute loss for weight  -0.616946  -0.616956 result 0.321847
 training batch 39 mu var00.327464
compute loss for weight  -0.616966  -0.616956 result 0.321869
 training batch 40 mu var00.327464
compute loss for weight  -0.616951  -0.616956 result 0.321852
 training batch 41 mu var00.327464
compute loss for weight  -0.616961  -0.616956 result 0.321863
   --dy = -1.12167 dy_ref = -1.12167
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m3.99736e-10[NON-XML-CHAR-0x1B][39m