Execution Time0.66s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: PR-4279-x86_64-fedora27-gcc7-opt (sft-fedora-27-2.cern.ch) on 2019-11-12 17:27:12
Repository revision: f842ddf766b5950c804aaefec33fd42ab33bc3b1

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 var00.526839
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.8486     -0.1701 
   1 |     0.7647     -0.3717 
   2 |    -0.4543      0.7035 
   3 |     0.9108     -0.4574 
   4 |     0.8605     -0.1946 
   5 |      1.986     -0.4802 
   6 |      1.405      -1.413 
   7 |     -2.315       1.578 
   8 |      1.144     -0.3887 
   9 |     0.1179      -0.131 

output BN 
output DL feature 0 mean 0.526839	output DL std 1.19734
output DL feature 1 mean -0.132533	output DL std 0.790171
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2833    -0.05014 
   1 |     0.2094      -0.319 
   2 |    -0.8637       1.115 
   3 |      0.338     -0.4334 
   4 |     0.2938    -0.08275 
   5 |      1.284     -0.4637 
   6 |     0.7729      -1.708 
   7 |     -2.501       2.282 
   8 |     0.5434     -0.3417 
   9 |      -0.36    0.002049 

output BN feature 0 mean -1.11022e-16	output BN std 1.05405
output BN feature 1 mean -3.42174e-17	output BN std 1.054
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.3994    -0.07588      0.4415     -0.4197 
   1 |     0.4678      0.4585        0.88      -1.153 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.6369      -0.971     -0.2482      0.5206 
   1 |    -0.1347      0.6526      0.2955       -0.57 

 training batch 2 mu var00.526842
compute loss for weight  0.636863  0.636853 result 0.605642
 training batch 3 mu var00.526839
compute loss for weight  0.636843  0.636853 result 0.605634
 training batch 4 mu var00.526839
compute loss for weight  0.636858  0.636853 result 0.60564
 training batch 5 mu var00.526839
compute loss for weight  0.636848  0.636853 result 0.605636
   --dy = 0.399444 dy_ref = 0.399444
 training batch 6 mu var00.526838
compute loss for weight  -0.970972  -0.970982 result 0.605638
 training batch 7 mu var00.526839
compute loss for weight  -0.970992  -0.970982 result 0.605639
 training batch 8 mu var00.526839
compute loss for weight  -0.970977  -0.970982 result 0.605638
 training batch 9 mu var00.526839
compute loss for weight  -0.970987  -0.970982 result 0.605639
   --dy = -0.0758799 dy_ref = -0.0758799
 training batch 10 mu var00.526839
compute loss for weight  -0.24821  -0.24822 result 0.605643
 training batch 11 mu var00.526839
compute loss for weight  -0.24823  -0.24822 result 0.605634
 training batch 12 mu var00.526839
compute loss for weight  -0.248215  -0.24822 result 0.605641
 training batch 13 mu var00.526839
compute loss for weight  -0.248225  -0.24822 result 0.605636
   --dy = 0.441515 dy_ref = 0.441515
 training batch 14 mu var00.526839
compute loss for weight  0.520646  0.520636 result 0.605634
 training batch 15 mu var00.526839
compute loss for weight  0.520626  0.520636 result 0.605643
 training batch 16 mu var00.526839
compute loss for weight  0.520641  0.520636 result 0.605636
 training batch 17 mu var00.526839
compute loss for weight  0.520631  0.520636 result 0.60564
   --dy = -0.41974 dy_ref = -0.41974
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.7612       1.973 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.526839
compute loss for weight  1.00001  1 result 0.605631
 training batch 19 mu var00.526839
compute loss for weight  0.99999  1 result 0.605646
 training batch 20 mu var00.526839
compute loss for weight  1.00001  1 result 0.605635
 training batch 21 mu var00.526839
compute loss for weight  0.999995  1 result 0.605642
   --dy = -0.761242 dy_ref = -0.761242
 training batch 22 mu var00.526839
compute loss for weight  1.00001  1 result 0.605658
 training batch 23 mu var00.526839
compute loss for weight  0.99999  1 result 0.605619
 training batch 24 mu var00.526839
compute loss for weight  1.00001  1 result 0.605648
 training batch 25 mu var00.526839
compute loss for weight  0.999995  1 result 0.605629
   --dy = 1.97252 dy_ref = 1.97252
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -1.596e-16  -1.804e-16 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.526839
compute loss for weight  1e-05  0 result 0.605638
 training batch 27 mu var00.526839
compute loss for weight  -1e-05  0 result 0.605638
 training batch 28 mu var00.526839
compute loss for weight  5e-06  0 result 0.605638
 training batch 29 mu var00.526839
compute loss for weight  -5e-06  0 result 0.605638
   --dy = 3.14563e-11 dy_ref = -1.59595e-16
 training batch 30 mu var00.526839
compute loss for weight  1e-05  0 result 0.605638
 training batch 31 mu var00.526839
compute loss for weight  -1e-05  0 result 0.605638
 training batch 32 mu var00.526839
compute loss for weight  5e-06  0 result 0.605638
 training batch 33 mu var00.526839
compute loss for weight  -5e-06  0 result 0.605638
   --dy = -1.66533e-11 dy_ref = -1.80411e-16
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.9811      -1.406 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.7759      -1.403 

 training batch 34 mu var00.526839
compute loss for weight  -0.775912  -0.775922 result 0.605648
 training batch 35 mu var00.526839
compute loss for weight  -0.775932  -0.775922 result 0.605629
 training batch 36 mu var00.526839
compute loss for weight  -0.775917  -0.775922 result 0.605643
 training batch 37 mu var00.526839
compute loss for weight  -0.775927  -0.775922 result 0.605633
   --dy = 0.981081 dy_ref = 0.981081
 training batch 38 mu var00.526839
compute loss for weight  -1.40329  -1.4033 result 0.605624
 training batch 39 mu var00.526839
compute loss for weight  -1.40331  -1.4033 result 0.605652
 training batch 40 mu var00.526839
compute loss for weight  -1.40329  -1.4033 result 0.605631
 training batch 41 mu var00.526839
compute loss for weight  -1.4033  -1.4033 result 0.605645
   --dy = -1.40563 dy_ref = -1.40563
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m1.24435e-10[NON-XML-CHAR-0x1B][39m