Execution Time0.25s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-centos7-gcc62-opt-no-rt-cxxmodules (olhswep09.cern.ch) on 2019-11-15 01:03:26
Repository revision: 14de58de35eff907054671888ccc2de0f7f27e77

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 var00.690431
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.148       1.834 
   1 |      1.274       1.531 
   2 |   -0.04537       1.286 
   3 |     0.4641       2.707 
   4 |      0.967        2.12 
   5 |      2.548       4.537 
   6 |    -0.5696       -1.95 
   7 |    -0.2073      -1.361 
   8 |      1.418      0.7739 
   9 |   -0.09196      0.1025 

output BN 
output DL feature 0 mean 0.690431	output DL std 0.954759
output DL feature 1 mean 1.15799	output DL std 1.90378
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.5052      0.3744 
   1 |     0.6439      0.2066 
   2 |    -0.8123     0.07073 
   3 |    -0.2499      0.8574 
   4 |     0.3053      0.5328 
   5 |      2.051       1.871 
   6 |     -1.391      -1.721 
   7 |    -0.9911      -1.395 
   8 |     0.8028     -0.2127 
   9 |    -0.8637     -0.5844 

output BN feature 0 mean -1.66533e-16	output BN std 1.05403
output BN feature 1 mean -2.22045e-17	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 |   -0.01771    -0.06711    -0.01467    -0.08864 
   1 |  -0.002133     0.05653    0.004059     0.06757 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |      1.303    0.006127     -0.2024     -0.2319 
   1 |      1.574     -0.7335     -0.1175      -1.153 

 training batch 2 mu var00.690433
compute loss for weight  1.30342  1.30341 result 0.390732
 training batch 3 mu var00.690431
compute loss for weight  1.3034  1.30341 result 0.390732
 training batch 4 mu var00.690431
compute loss for weight  1.30341  1.30341 result 0.390732
 training batch 5 mu var00.690431
compute loss for weight  1.3034  1.30341 result 0.390732
   --dy = -0.0177138 dy_ref = -0.0177138
 training batch 6 mu var00.69043
compute loss for weight  0.00613653  0.00612653 result 0.390731
 training batch 7 mu var00.690431
compute loss for weight  0.00611653  0.00612653 result 0.390733
 training batch 8 mu var00.69043
compute loss for weight  0.00613153  0.00612653 result 0.390732
 training batch 9 mu var00.690431
compute loss for weight  0.00612153  0.00612653 result 0.390732
   --dy = -0.0671138 dy_ref = -0.0671138
 training batch 10 mu var00.690431
compute loss for weight  -0.202435  -0.202445 result 0.390732
 training batch 11 mu var00.690431
compute loss for weight  -0.202455  -0.202445 result 0.390732
 training batch 12 mu var00.690431
compute loss for weight  -0.20244  -0.202445 result 0.390732
 training batch 13 mu var00.690431
compute loss for weight  -0.20245  -0.202445 result 0.390732
   --dy = -0.014674 dy_ref = -0.014674
 training batch 14 mu var00.690431
compute loss for weight  -0.231929  -0.231939 result 0.390731
 training batch 15 mu var00.690431
compute loss for weight  -0.231949  -0.231939 result 0.390733
 training batch 16 mu var00.690431
compute loss for weight  -0.231934  -0.231939 result 0.390732
 training batch 17 mu var00.690431
compute loss for weight  -0.231944  -0.231939 result 0.390732
   --dy = -0.0886383 dy_ref = -0.0886383
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2457      0.5358 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.690431
compute loss for weight  1.00001  1 result 0.390734
 training batch 19 mu var00.690431
compute loss for weight  0.99999  1 result 0.39073
 training batch 20 mu var00.690431
compute loss for weight  1.00001  1 result 0.390733
 training batch 21 mu var00.690431
compute loss for weight  0.999995  1 result 0.390731
   --dy = 0.245672 dy_ref = 0.245672
 training batch 22 mu var00.690431
compute loss for weight  1.00001  1 result 0.390737
 training batch 23 mu var00.690431
compute loss for weight  0.99999  1 result 0.390727
 training batch 24 mu var00.690431
compute loss for weight  1.00001  1 result 0.390735
 training batch 25 mu var00.690431
compute loss for weight  0.999995  1 result 0.390729
   --dy = 0.535792 dy_ref = 0.535792
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -1.735e-17  -4.857e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.690431
compute loss for weight  1e-05  0 result 0.390732
 training batch 27 mu var00.690431
compute loss for weight  -1e-05  0 result 0.390732
 training batch 28 mu var00.690431
compute loss for weight  5e-06  0 result 0.390732
 training batch 29 mu var00.690431
compute loss for weight  -5e-06  0 result 0.390732
   --dy = -8.32667e-12 dy_ref = -1.73472e-17
 training batch 30 mu var00.690431
compute loss for weight  1e-05  0 result 0.390732
 training batch 31 mu var00.690431
compute loss for weight  -1e-05  0 result 0.390732
 training batch 32 mu var00.690431
compute loss for weight  5e-06  0 result 0.390732
 training batch 33 mu var00.690431
compute loss for weight  -5e-06  0 result 0.390732
   --dy = -8.32667e-12 dy_ref = -4.85723e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.143       1.225 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2149      0.4373 

 training batch 34 mu var00.690431
compute loss for weight  0.214884  0.214874 result 0.390743
 training batch 35 mu var00.690431
compute loss for weight  0.214864  0.214874 result 0.390721
 training batch 36 mu var00.690431
compute loss for weight  0.214879  0.214874 result 0.390738
 training batch 37 mu var00.690431
compute loss for weight  0.214869  0.214874 result 0.390726
   --dy = 1.14333 dy_ref = 1.14333
 training batch 38 mu var00.690431
compute loss for weight  0.43731  0.4373 result 0.390744
 training batch 39 mu var00.690431
compute loss for weight  0.43729  0.4373 result 0.39072
 training batch 40 mu var00.690431
compute loss for weight  0.437305  0.4373 result 0.390738
 training batch 41 mu var00.690431
compute loss for weight  0.437295  0.4373 result 0.390726
   --dy = 1.22523 dy_ref = 1.22523
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m6.59808e-10[NON-XML-CHAR-0x1B][39m