Execution Time0.05s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: PR-420-x86_64-ubuntu16-gcc54-opt (sft-ubuntu-1604-4) on 2019-11-14 09:36:08
Repository revision: 95d28fe99be54b4de35e2d05610021f3b746a746

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 var00.356673
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.4769    -0.08113 
   1 |     0.9972      -1.734 
   2 |     0.1384       2.345 
   3 |      1.151      0.5956 
   4 |     0.5819      0.3498 
   5 |      1.377     -0.1931 
   6 |     -1.307       -1.28 
   7 |     0.2476     -0.1763 
   8 |    -0.1599      -1.072 
   9 |    0.06272     0.01341 

output BN 
output DL feature 0 mean 0.356673	output DL std 0.769367
output DL feature 1 mean -0.123378	output DL std 1.13983
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1647     0.03907 
   1 |     0.8774       -1.49 
   2 |    -0.2991       2.282 
   3 |      1.088      0.6649 
   4 |     0.3085      0.4376 
   5 |      1.398    -0.06448 
   6 |     -2.278       -1.07 
   7 |    -0.1494    -0.04898 
   8 |    -0.7077     -0.8772 
   9 |    -0.4027      0.1265 

output BN feature 0 mean -1.11022e-17	output BN std 1.05399
output BN feature 1 mean 1.66533e-17	output BN std 1.05405
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.2147     0.09688     -0.2313      0.3816 
   1 |    -0.1791        0.13     0.06356      0.3327 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |      0.768     0.07738     -0.3539     -0.6663 
   1 |     -1.077     -0.3858       1.028     -0.7756 

 training batch 2 mu var00.356676
compute loss for weight  0.767965  0.767955 result 0.88291
 training batch 3 mu var00.356673
compute loss for weight  0.767945  0.767955 result 0.882906
 training batch 4 mu var00.356674
compute loss for weight  0.76796  0.767955 result 0.882909
 training batch 5 mu var00.356673
compute loss for weight  0.76795  0.767955 result 0.882907
   --dy = 0.214734 dy_ref = 0.214734
 training batch 6 mu var00.356673
compute loss for weight  0.077393  0.077383 result 0.882909
 training batch 7 mu var00.356673
compute loss for weight  0.077373  0.077383 result 0.882907
 training batch 8 mu var00.356673
compute loss for weight  0.077388  0.077383 result 0.882909
 training batch 9 mu var00.356673
compute loss for weight  0.077378  0.077383 result 0.882908
   --dy = 0.0968768 dy_ref = 0.0968768
 training batch 10 mu var00.356674
compute loss for weight  -0.35385  -0.35386 result 0.882906
 training batch 11 mu var00.356673
compute loss for weight  -0.35387  -0.35386 result 0.882911
 training batch 12 mu var00.356673
compute loss for weight  -0.353855  -0.35386 result 0.882907
 training batch 13 mu var00.356673
compute loss for weight  -0.353865  -0.35386 result 0.882909
   --dy = -0.231319 dy_ref = -0.231319
 training batch 14 mu var00.356673
compute loss for weight  -0.666307  -0.666317 result 0.882912
 training batch 15 mu var00.356673
compute loss for weight  -0.666327  -0.666317 result 0.882904
 training batch 16 mu var00.356673
compute loss for weight  -0.666312  -0.666317 result 0.88291
 training batch 17 mu var00.356673
compute loss for weight  -0.666322  -0.666317 result 0.882906
   --dy = 0.381566 dy_ref = 0.381566
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    0.07409       1.692 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.356673
compute loss for weight  1.00001  1 result 0.882909
 training batch 19 mu var00.356673
compute loss for weight  0.99999  1 result 0.882908
 training batch 20 mu var00.356673
compute loss for weight  1.00001  1 result 0.882909
 training batch 21 mu var00.356673
compute loss for weight  0.999995  1 result 0.882908
   --dy = 0.0740877 dy_ref = 0.0740877
 training batch 22 mu var00.356673
compute loss for weight  1.00001  1 result 0.882925
 training batch 23 mu var00.356673
compute loss for weight  0.99999  1 result 0.882891
 training batch 24 mu var00.356673
compute loss for weight  1.00001  1 result 0.882917
 training batch 25 mu var00.356673
compute loss for weight  0.999995  1 result 0.8829
   --dy = 1.69173 dy_ref = 1.69173
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  1.735e-18  -6.939e-18 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.356673
compute loss for weight  1e-05  0 result 0.882908
 training batch 27 mu var00.356673
compute loss for weight  -1e-05  0 result 0.882908
 training batch 28 mu var00.356673
compute loss for weight  5e-06  0 result 0.882908
 training batch 29 mu var00.356673
compute loss for weight  -5e-06  0 result 0.882908
   --dy = -3.88578e-11 dy_ref = 1.73472e-18
 training batch 30 mu var00.356673
compute loss for weight  1e-05  0 result 0.882908
 training batch 31 mu var00.356673
compute loss for weight  -1e-05  0 result 0.882908
 training batch 32 mu var00.356673
compute loss for weight  5e-06  0 result 0.882908
 training batch 33 mu var00.356673
compute loss for weight  -5e-06  0 result 0.882908
   --dy = -7.40149e-12 dy_ref = -6.93889e-18
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.2506       1.787 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.2957      0.9467 

 training batch 34 mu var00.356673
compute loss for weight  -0.29568  -0.29569 result 0.882906
 training batch 35 mu var00.356673
compute loss for weight  -0.2957  -0.29569 result 0.882911
 training batch 36 mu var00.356673
compute loss for weight  -0.295685  -0.29569 result 0.882907
 training batch 37 mu var00.356673
compute loss for weight  -0.295695  -0.29569 result 0.88291
   --dy = -0.250559 dy_ref = -0.250559
 training batch 38 mu var00.356673
compute loss for weight  0.946748  0.946738 result 0.882926
 training batch 39 mu var00.356673
compute loss for weight  0.946728  0.946738 result 0.88289
 training batch 40 mu var00.356673
compute loss for weight  0.946743  0.946738 result 0.882917
 training batch 41 mu var00.356673
compute loss for weight  0.946733  0.946738 result 0.882899
   --dy = 1.7869 dy_ref = 1.7869
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m5.15788e-10[NON-XML-CHAR-0x1B][39m