Execution Time0.21s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-centos7-gcc62-opt-no-rt-cxxmodules (olsnba08.cern.ch) on 2019-11-13 01:18:50
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 var0-0.344404
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.7605      0.2828 
   1 |    -0.1389       1.782 
   2 |    0.00388     -0.8656 
   3 |       1.29       2.033 
   4 |    -0.3398      0.5007 
   5 |     -1.177       1.308 
   6 |    -0.4871     -0.7034 
   7 |     -0.142      -0.498 
   8 |      -1.97     -0.6661 
   9 |     0.2775      0.2713 

output BN 
output DL feature 0 mean -0.344404	output DL std 0.868681
output DL feature 1 mean 0.34442	output DL std 1.06312
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.5049     -0.0611 
   1 |     0.2493       1.425 
   2 |     0.4226        -1.2 
   3 |      1.983       1.674 
   4 |   0.005622       0.155 
   5 |     -1.011      0.9552 
   6 |    -0.1731      -1.039 
   7 |     0.2456     -0.8353 
   8 |     -1.972      -1.002 
   9 |     0.7546    -0.07245 

output BN feature 0 mean -1.11022e-17	output BN std 1.05401
output BN feature 1 mean 1.38778e-18	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 |    -0.5821        0.54      0.3021      0.2997 
   1 |     0.6865     0.02251      0.5883       0.501 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     -0.601     -0.2289     -0.3477     -0.4044 
   1 |     0.9761     -0.2284      -1.031     -0.4739 

 training batch 2 mu var0-0.344401
compute loss for weight  -0.600957  -0.600967 result 0.805278
 training batch 3 mu var0-0.344404
compute loss for weight  -0.600977  -0.600967 result 0.80529
 training batch 4 mu var0-0.344403
compute loss for weight  -0.600962  -0.600967 result 0.805281
 training batch 5 mu var0-0.344404
compute loss for weight  -0.600972  -0.600967 result 0.805287
   --dy = -0.582145 dy_ref = -0.582145
 training batch 6 mu var0-0.344404
compute loss for weight  -0.228853  -0.228863 result 0.805289
 training batch 7 mu var0-0.344404
compute loss for weight  -0.228873  -0.228863 result 0.805279
 training batch 8 mu var0-0.344404
compute loss for weight  -0.228858  -0.228863 result 0.805287
 training batch 9 mu var0-0.344404
compute loss for weight  -0.228868  -0.228863 result 0.805281
   --dy = 0.539998 dy_ref = 0.539998
 training batch 10 mu var0-0.344404
compute loss for weight  -0.347698  -0.347708 result 0.805287
 training batch 11 mu var0-0.344404
compute loss for weight  -0.347718  -0.347708 result 0.805281
 training batch 12 mu var0-0.344404
compute loss for weight  -0.347703  -0.347708 result 0.805285
 training batch 13 mu var0-0.344404
compute loss for weight  -0.347713  -0.347708 result 0.805282
   --dy = 0.302083 dy_ref = 0.302083
 training batch 14 mu var0-0.344404
compute loss for weight  -0.404386  -0.404396 result 0.805287
 training batch 15 mu var0-0.344404
compute loss for weight  -0.404406  -0.404396 result 0.805281
 training batch 16 mu var0-0.344404
compute loss for weight  -0.404391  -0.404396 result 0.805285
 training batch 17 mu var0-0.344404
compute loss for weight  -0.404401  -0.404396 result 0.805282
   --dy = 0.299695 dy_ref = 0.299695
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2191       1.391 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.344404
compute loss for weight  1.00001  1 result 0.805286
 training batch 19 mu var0-0.344404
compute loss for weight  0.99999  1 result 0.805282
 training batch 20 mu var0-0.344404
compute loss for weight  1.00001  1 result 0.805285
 training batch 21 mu var0-0.344404
compute loss for weight  0.999995  1 result 0.805283
   --dy = 0.219125 dy_ref = 0.219125
 training batch 22 mu var0-0.344404
compute loss for weight  1.00001  1 result 0.805298
 training batch 23 mu var0-0.344404
compute loss for weight  0.99999  1 result 0.80527
 training batch 24 mu var0-0.344404
compute loss for weight  1.00001  1 result 0.805291
 training batch 25 mu var0-0.344404
compute loss for weight  0.999995  1 result 0.805277
   --dy = 1.39144 dy_ref = 1.39144
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -7.633e-17   2.776e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.344404
compute loss for weight  1e-05  0 result 0.805284
 training batch 27 mu var0-0.344404
compute loss for weight  -1e-05  0 result 0.805284
 training batch 28 mu var0-0.344404
compute loss for weight  5e-06  0 result 0.805284
 training batch 29 mu var0-0.344404
compute loss for weight  -5e-06  0 result 0.805284
   --dy = 1.85037e-12 dy_ref = -7.63278e-17
 training batch 30 mu var0-0.344404
compute loss for weight  1e-05  0 result 0.805284
 training batch 31 mu var0-0.344404
compute loss for weight  -1e-05  0 result 0.805284
 training batch 32 mu var0-0.344404
compute loss for weight  5e-06  0 result 0.805284
 training batch 33 mu var0-0.344404
compute loss for weight  -5e-06  0 result 0.805284
   --dy = 1.85037e-11 dy_ref = 2.77556e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.3743      -1.444 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.5853     -0.9638 

 training batch 34 mu var0-0.344404
compute loss for weight  0.585357  0.585347 result 0.805288
 training batch 35 mu var0-0.344404
compute loss for weight  0.585337  0.585347 result 0.80528
 training batch 36 mu var0-0.344404
compute loss for weight  0.585352  0.585347 result 0.805286
 training batch 37 mu var0-0.344404
compute loss for weight  0.585342  0.585347 result 0.805282
   --dy = 0.37435 dy_ref = 0.37435
 training batch 38 mu var0-0.344404
compute loss for weight  -0.963749  -0.963759 result 0.805269
 training batch 39 mu var0-0.344404
compute loss for weight  -0.963769  -0.963759 result 0.805298
 training batch 40 mu var0-0.344404
compute loss for weight  -0.963754  -0.963759 result 0.805277
 training batch 41 mu var0-0.344404
compute loss for weight  -0.963764  -0.963759 result 0.805291
   --dy = -1.44377 dy_ref = -1.44377
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m1.36362e-10[NON-XML-CHAR-0x1B][39m