Execution Time0.15s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-centos7-gcc62-opt-no-rt-cxxmodules (olsnba08.cern.ch) on 2019-11-13 12:58:27
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 var0-0.622425
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.304    0.002409 
   1 |      1.364      0.3375 
   2 |     -2.516      -1.794 
   3 |    -0.7444      -2.384 
   4 |     -1.563     -0.6675 
   5 |     -2.606      -0.607 
   6 |    -0.2328       2.338 
   7 |      2.552      0.1784 
   8 |     -1.234       2.036 
   9 |    0.06014     -0.2487 

output BN 
output DL feature 0 mean -0.622425	output DL std 1.6297
output DL feature 1 mean -0.0808832	output DL std 1.4709
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.4407     0.05969 
   1 |      1.285      0.2998 
   2 |     -1.225      -1.228 
   3 |   -0.07889       -1.65 
   4 |    -0.6085     -0.4204 
   5 |     -1.283      -0.377 
   6 |      0.252       1.734 
   7 |      2.053      0.1858 
   8 |    -0.3955       1.517 
   9 |     0.4415     -0.1203 

output BN feature 0 mean -3.88578e-17	output BN std 1.05407
output BN feature 1 mean 8.74301e-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 |    -0.3055      -0.173     -0.3567      -1.066 
   1 |     0.6017     -0.7163      0.8746      0.3979 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.3178       1.348      -1.319      0.1316 
   1 |     0.3707      0.4295     -0.1781       1.394 

 training batch 2 mu var0-0.622422
compute loss for weight  0.317782  0.317772 result 6.31502
 training batch 3 mu var0-0.622425
compute loss for weight  0.317762  0.317772 result 6.31502
 training batch 4 mu var0-0.622424
compute loss for weight  0.317777  0.317772 result 6.31502
 training batch 5 mu var0-0.622425
compute loss for weight  0.317767  0.317772 result 6.31502
   --dy = -0.305484 dy_ref = -0.305484
 training batch 6 mu var0-0.622425
compute loss for weight  1.34781  1.3478 result 6.31502
 training batch 7 mu var0-0.622425
compute loss for weight  1.34779  1.3478 result 6.31502
 training batch 8 mu var0-0.622425
compute loss for weight  1.3478  1.3478 result 6.31502
 training batch 9 mu var0-0.622425
compute loss for weight  1.34779  1.3478 result 6.31502
   --dy = -0.173005 dy_ref = -0.173005
 training batch 10 mu var0-0.622425
compute loss for weight  -1.31919  -1.3192 result 6.31502
 training batch 11 mu var0-0.622425
compute loss for weight  -1.31921  -1.3192 result 6.31502
 training batch 12 mu var0-0.622425
compute loss for weight  -1.31919  -1.3192 result 6.31502
 training batch 13 mu var0-0.622425
compute loss for weight  -1.3192  -1.3192 result 6.31502
   --dy = -0.356734 dy_ref = -0.356734
 training batch 14 mu var0-0.622425
compute loss for weight  0.131649  0.131639 result 6.31501
 training batch 15 mu var0-0.622425
compute loss for weight  0.131629  0.131639 result 6.31503
 training batch 16 mu var0-0.622425
compute loss for weight  0.131644  0.131639 result 6.31502
 training batch 17 mu var0-0.622425
compute loss for weight  0.131634  0.131639 result 6.31503
   --dy = -1.06628 dy_ref = -1.06628
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -0.283       12.91 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.622425
compute loss for weight  1.00001  1 result 6.31502
 training batch 19 mu var0-0.622425
compute loss for weight  0.99999  1 result 6.31502
 training batch 20 mu var0-0.622425
compute loss for weight  1.00001  1 result 6.31502
 training batch 21 mu var0-0.622425
compute loss for weight  0.999995  1 result 6.31502
   --dy = -0.282958 dy_ref = -0.282958
 training batch 22 mu var0-0.622425
compute loss for weight  1.00001  1 result 6.31515
 training batch 23 mu var0-0.622425
compute loss for weight  0.99999  1 result 6.31489
 training batch 24 mu var0-0.622425
compute loss for weight  1.00001  1 result 6.31509
 training batch 25 mu var0-0.622425
compute loss for weight  0.999995  1 result 6.31496
   --dy = 12.913 dy_ref = 12.913
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  -1.11e-16   1.471e-15 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.622425
compute loss for weight  1e-05  0 result 6.31502
 training batch 27 mu var0-0.622425
compute loss for weight  -1e-05  0 result 6.31502
 training batch 28 mu var0-0.622425
compute loss for weight  5e-06  0 result 6.31502
 training batch 29 mu var0-0.622425
compute loss for weight  -5e-06  0 result 6.31502
   --dy = -1.33227e-10 dy_ref = -1.11022e-16
 training batch 30 mu var0-0.622425
compute loss for weight  1e-05  0 result 6.31502
 training batch 31 mu var0-0.622425
compute loss for weight  -1e-05  0 result 6.31502
 training batch 32 mu var0-0.622425
compute loss for weight  5e-06  0 result 6.31502
 training batch 33 mu var0-0.622425
compute loss for weight  -5e-06  0 result 6.31502
   --dy = 0 dy_ref = 1.47105e-15
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.7543      -4.974 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.3751      -2.596 

 training batch 34 mu var0-0.622425
compute loss for weight  0.375121  0.375111 result 6.31501
 training batch 35 mu var0-0.622425
compute loss for weight  0.375101  0.375111 result 6.31503
 training batch 36 mu var0-0.622425
compute loss for weight  0.375116  0.375111 result 6.31502
 training batch 37 mu var0-0.622425
compute loss for weight  0.375106  0.375111 result 6.31502
   --dy = -0.754331 dy_ref = -0.754331
 training batch 38 mu var0-0.622425
compute loss for weight  -2.59595  -2.59596 result 6.31497
 training batch 39 mu var0-0.622425
compute loss for weight  -2.59597  -2.59596 result 6.31507
 training batch 40 mu var0-0.622425
compute loss for weight  -2.59596  -2.59596 result 6.315
 training batch 41 mu var0-0.622425
compute loss for weight  -2.59597  -2.59596 result 6.31505
   --dy = -4.97426 dy_ref = -4.97426
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m2.71678e-10[NON-XML-CHAR-0x1B][39m