Execution Time0.31s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-mac1014-clang100-opt-exp-pyroot (macphsft17.dyndns.cern.ch) on 2019-11-15 03:48:45

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 var00.908164
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.707     -0.1865 
   1 |     0.9776    -0.04038 
   2 |    -0.5909       1.394 
   3 |    -0.7942       2.482 
   4 |      1.173       0.482 
   5 |      3.256      0.3048 
   6 |      2.006      -2.415 
   7 |     -1.958     0.09655 
   8 |      3.535      -2.322 
   9 |    -0.2314      0.2885 

output BN 
output DL feature 0 mean 0.908164	output DL std 1.79489
output DL feature 1 mean 0.00835153	output DL std 1.48264
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.4693     -0.1385 
   1 |     0.0408    -0.03464 
   2 |    -0.8803      0.9853 
   3 |    -0.9997       1.759 
   4 |     0.1558      0.3367 
   5 |      1.379      0.2107 
   6 |     0.6449      -1.723 
   7 |     -1.683      0.0627 
   8 |      1.543      -1.657 
   9 |    -0.6692      0.1992 

output BN feature 0 mean 0	output BN std 1.05407
output BN feature 1 mean -5.55112e-18	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 |   -0.05672      0.1364   -0.007852       0.154 
   1 |    -0.1132      0.2595    -0.01944      0.2823 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |      1.381     -0.5929      0.1294       1.041 
   1 |    -0.2884      -0.286    -0.08358      -1.416 

 training batch 2 mu var00.908166
compute loss for weight  1.38107  1.38106 result 2.08949
 training batch 3 mu var00.908164
compute loss for weight  1.38105  1.38106 result 2.08949
 training batch 4 mu var00.908164
compute loss for weight  1.38106  1.38106 result 2.08949
 training batch 5 mu var00.908164
compute loss for weight  1.38105  1.38106 result 2.08949
   --dy = -0.0567195 dy_ref = -0.0567195
 training batch 6 mu var00.908163
compute loss for weight  -0.592895  -0.592905 result 2.08949
 training batch 7 mu var00.908164
compute loss for weight  -0.592915  -0.592905 result 2.08949
 training batch 8 mu var00.908163
compute loss for weight  -0.5929  -0.592905 result 2.08949
 training batch 9 mu var00.908164
compute loss for weight  -0.59291  -0.592905 result 2.08949
   --dy = 0.136431 dy_ref = 0.136431
 training batch 10 mu var00.908164
compute loss for weight  0.129414  0.129404 result 2.08949
 training batch 11 mu var00.908164
compute loss for weight  0.129394  0.129404 result 2.08949
 training batch 12 mu var00.908164
compute loss for weight  0.129409  0.129404 result 2.08949
 training batch 13 mu var00.908164
compute loss for weight  0.129399  0.129404 result 2.08949
   --dy = -0.00785237 dy_ref = -0.00785237
 training batch 14 mu var00.908163
compute loss for weight  1.04094  1.04093 result 2.08949
 training batch 15 mu var00.908164
compute loss for weight  1.04092  1.04093 result 2.08949
 training batch 16 mu var00.908163
compute loss for weight  1.04093  1.04093 result 2.08949
 training batch 17 mu var00.908164
compute loss for weight  1.04092  1.04093 result 2.08949
   --dy = 0.153952 dy_ref = 0.153952
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.3635       3.816 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.908164
compute loss for weight  1.00001  1 result 2.08949
 training batch 19 mu var00.908164
compute loss for weight  0.99999  1 result 2.08948
 training batch 20 mu var00.908164
compute loss for weight  1.00001  1 result 2.08949
 training batch 21 mu var00.908164
compute loss for weight  0.999995  1 result 2.08948
   --dy = 0.363468 dy_ref = 0.363468
 training batch 22 mu var00.908164
compute loss for weight  1.00001  1 result 2.08952
 training batch 23 mu var00.908164
compute loss for weight  0.99999  1 result 2.08945
 training batch 24 mu var00.908164
compute loss for weight  1.00001  1 result 2.08951
 training batch 25 mu var00.908164
compute loss for weight  0.999995  1 result 2.08947
   --dy = 3.8155 dy_ref = 3.8155
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -1.214e-17  -1.388e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.908164
compute loss for weight  1e-05  0 result 2.08949
 training batch 27 mu var00.908164
compute loss for weight  -1e-05  0 result 2.08949
 training batch 28 mu var00.908164
compute loss for weight  5e-06  0 result 2.08949
 training batch 29 mu var00.908164
compute loss for weight  -5e-06  0 result 2.08949
   --dy = 5.92119e-11 dy_ref = -1.21431e-17
 training batch 30 mu var00.908164
compute loss for weight  1e-05  0 result 2.08949
 training batch 31 mu var00.908164
compute loss for weight  -1e-05  0 result 2.08949
 training batch 32 mu var00.908164
compute loss for weight  5e-06  0 result 2.08949
 training batch 33 mu var00.908164
compute loss for weight  -5e-06  0 result 2.08949
   --dy = 5.92119e-11 dy_ref = -1.38778e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -2.019       2.877 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      -0.18       1.326 

 training batch 34 mu var00.908164
compute loss for weight  -0.180014  -0.180024 result 2.08947
 training batch 35 mu var00.908164
compute loss for weight  -0.180034  -0.180024 result 2.08951
 training batch 36 mu var00.908164
compute loss for weight  -0.180019  -0.180024 result 2.08948
 training batch 37 mu var00.908164
compute loss for weight  -0.180029  -0.180024 result 2.0895
   --dy = -2.019 dy_ref = -2.019
 training batch 38 mu var00.908164
compute loss for weight  1.3261  1.32609 result 2.08952
 training batch 39 mu var00.908164
compute loss for weight  1.32608  1.32609 result 2.08946
 training batch 40 mu var00.908164
compute loss for weight  1.3261  1.32609 result 2.0895
 training batch 41 mu var00.908164
compute loss for weight  1.32609  1.32609 result 2.08947
   --dy = 2.87726 dy_ref = 2.87726
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m3.41797e-09[NON-XML-CHAR-0x1B][39m