Execution Time0.19s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-centos7-gcc62-opt-master (olhswep09.cern.ch) on 2019-11-14 11:14:22
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 var00.610006
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.307      0.1971 
   1 |     -0.321      0.4805 
   2 |      1.004      0.1307 
   3 |     -1.023      0.3452 
   4 |     0.9695       0.203 
   5 |      2.274      0.5492 
   6 |       0.39     -0.9726 
   7 |    -0.6655      0.6239 
   8 |      2.484    -0.08673 
   9 |    -0.3189    -0.01715 

output BN 
output DL feature 0 mean 0.610006	output DL std 1.20754
output DL feature 1 mean 0.145317	output DL std 0.45713
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.6086      0.1194 
   1 |    -0.8127      0.7727 
   2 |      0.344    -0.03359 
   3 |     -1.426      0.4607 
   4 |     0.3138      0.1329 
   5 |      1.453      0.9312 
   6 |     -0.192      -2.577 
   7 |     -1.113       1.103 
   8 |      1.635     -0.5349 
   9 |    -0.8108     -0.3745 

output BN feature 0 mean 3.33067e-17	output BN std 1.05405
output BN feature 1 mean -1.11022e-17	output BN std 1.05381
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    0.05272     0.02439     0.01304     -0.1263 
   1 |    0.05685     -0.2954      0.4206       1.017 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.5567     -0.2083      0.8556      0.2803 
   1 |     0.4102      0.2701     -0.1387     -0.4379 

 training batch 2 mu var00.610009
compute loss for weight  0.55674  0.55673 result 0.194596
 training batch 3 mu var00.610006
compute loss for weight  0.55672  0.55673 result 0.194595
 training batch 4 mu var00.610007
compute loss for weight  0.556735  0.55673 result 0.194596
 training batch 5 mu var00.610006
compute loss for weight  0.556725  0.55673 result 0.194595
   --dy = 0.052722 dy_ref = 0.052722
 training batch 6 mu var00.610006
compute loss for weight  -0.208335  -0.208345 result 0.194596
 training batch 7 mu var00.610006
compute loss for weight  -0.208355  -0.208345 result 0.194595
 training batch 8 mu var00.610006
compute loss for weight  -0.20834  -0.208345 result 0.194596
 training batch 9 mu var00.610006
compute loss for weight  -0.20835  -0.208345 result 0.194595
   --dy = 0.0243945 dy_ref = 0.0243945
 training batch 10 mu var00.610006
compute loss for weight  0.855605  0.855595 result 0.194596
 training batch 11 mu var00.610006
compute loss for weight  0.855585  0.855595 result 0.194595
 training batch 12 mu var00.610006
compute loss for weight  0.8556  0.855595 result 0.194596
 training batch 13 mu var00.610006
compute loss for weight  0.85559  0.855595 result 0.194595
   --dy = 0.01304 dy_ref = 0.01304
 training batch 14 mu var00.610006
compute loss for weight  0.280354  0.280344 result 0.194594
 training batch 15 mu var00.610006
compute loss for weight  0.280334  0.280344 result 0.194597
 training batch 16 mu var00.610006
compute loss for weight  0.280349  0.280344 result 0.194595
 training batch 17 mu var00.610006
compute loss for weight  0.280339  0.280344 result 0.194596
   --dy = -0.12631 dy_ref = -0.12631
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2122       0.177 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.610006
compute loss for weight  1.00001  1 result 0.194598
 training batch 19 mu var00.610006
compute loss for weight  0.99999  1 result 0.194593
 training batch 20 mu var00.610006
compute loss for weight  1.00001  1 result 0.194597
 training batch 21 mu var00.610006
compute loss for weight  0.999995  1 result 0.194594
   --dy = 0.212187 dy_ref = 0.212187
 training batch 22 mu var00.610006
compute loss for weight  1.00001  1 result 0.194597
 training batch 23 mu var00.610006
compute loss for weight  0.99999  1 result 0.194594
 training batch 24 mu var00.610006
compute loss for weight  1.00001  1 result 0.194596
 training batch 25 mu var00.610006
compute loss for weight  0.999995  1 result 0.194595
   --dy = 0.177004 dy_ref = 0.177004
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

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

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.610006
compute loss for weight  1e-05  0 result 0.194595
 training batch 27 mu var00.610006
compute loss for weight  -1e-05  0 result 0.194595
 training batch 28 mu var00.610006
compute loss for weight  5e-06  0 result 0.194595
 training batch 29 mu var00.610006
compute loss for weight  -5e-06  0 result 0.194595
   --dy = -6.93889e-12 dy_ref = 1.38778e-17
 training batch 30 mu var00.610006
compute loss for weight  1e-05  0 result 0.194595
 training batch 31 mu var00.610006
compute loss for weight  -1e-05  0 result 0.194595
 training batch 32 mu var00.610006
compute loss for weight  5e-06  0 result 0.194595
 training batch 33 mu var00.610006
compute loss for weight  -5e-06  0 result 0.194595
   --dy = 4.62593e-13 dy_ref = 1.38778e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.6164     -0.5571 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.3442     -0.3177 

 training batch 34 mu var00.610006
compute loss for weight  -0.344234  -0.344244 result 0.194589
 training batch 35 mu var00.610006
compute loss for weight  -0.344254  -0.344244 result 0.194602
 training batch 36 mu var00.610006
compute loss for weight  -0.344239  -0.344244 result 0.194592
 training batch 37 mu var00.610006
compute loss for weight  -0.344249  -0.344244 result 0.194599
   --dy = -0.616385 dy_ref = -0.616385
 training batch 38 mu var00.610006
compute loss for weight  -0.317727  -0.317737 result 0.19459
 training batch 39 mu var00.610006
compute loss for weight  -0.317747  -0.317737 result 0.194601
 training batch 40 mu var00.610006
compute loss for weight  -0.317732  -0.317737 result 0.194593
 training batch 41 mu var00.610006
compute loss for weight  -0.317742  -0.317737 result 0.194598
   --dy = -0.557076 dy_ref = -0.557076
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m3.2855e-10[NON-XML-CHAR-0x1B][39m