Execution Time0.51s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: PR-4624-x86_64-fedora30-gcc9-opt (root-fedora30-1.cern.ch) on 2019-11-14 19:01:46

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 var0-0.272543
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      0.102      0.1214 
   1 |     -1.719       -0.81 
   2 |     0.4333     -0.1018 
   3 |     -4.982      -1.564 
   4 |    -0.8954     -0.1774 
   5 |     -1.308     -0.2859 
   6 |     0.7869       1.722 
   7 |      2.682     -0.6429 
   8 |       2.96       1.379 
   9 |    -0.7852     -0.1828 

output BN 
output DL feature 0 mean -0.272543	output DL std 2.28365
output DL feature 1 mean -0.0542965	output DL std 0.973275
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1729      0.1902 
   1 |    -0.6676     -0.8185 
   2 |     0.3258    -0.05147 
   3 |     -2.174      -1.635 
   4 |    -0.2875     -0.1333 
   5 |     -0.478     -0.2509 
   6 |      0.489       1.923 
   7 |      1.364     -0.6374 
   8 |      1.492       1.553 
   9 |    -0.2366     -0.1391 

output BN feature 0 mean -5.55112e-17	output BN std 1.05408
output BN feature 1 mean -3.60822e-17	output BN std 1.05403
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |   -0.06438      0.2389    -0.05697     -0.3141 
   1 |    -0.9962      -1.498      -3.834      -6.773 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    -0.3347         1.5       1.429      0.9498 
   1 |    -0.3369     -0.1708      0.5259      0.9145 

 training batch 2 mu var0-0.27254
compute loss for weight  -0.334707  -0.334717 result 1.38803
 training batch 3 mu var0-0.272543
compute loss for weight  -0.334727  -0.334717 result 1.38803
 training batch 4 mu var0-0.272542
compute loss for weight  -0.334712  -0.334717 result 1.38803
 training batch 5 mu var0-0.272543
compute loss for weight  -0.334722  -0.334717 result 1.38803
   --dy = -0.0643759 dy_ref = -0.0643759
 training batch 6 mu var0-0.272544
compute loss for weight  1.49967  1.49966 result 1.38803
 training batch 7 mu var0-0.272543
compute loss for weight  1.49965  1.49966 result 1.38802
 training batch 8 mu var0-0.272543
compute loss for weight  1.49967  1.49966 result 1.38803
 training batch 9 mu var0-0.272543
compute loss for weight  1.49966  1.49966 result 1.38802
   --dy = 0.238907 dy_ref = 0.238907
 training batch 10 mu var0-0.272543
compute loss for weight  1.42867  1.42866 result 1.38803
 training batch 11 mu var0-0.272543
compute loss for weight  1.42865  1.42866 result 1.38803
 training batch 12 mu var0-0.272543
compute loss for weight  1.42867  1.42866 result 1.38803
 training batch 13 mu var0-0.272543
compute loss for weight  1.42866  1.42866 result 1.38803
   --dy = -0.0569673 dy_ref = -0.0569673
 training batch 14 mu var0-0.272543
compute loss for weight  0.949797  0.949787 result 1.38802
 training batch 15 mu var0-0.272543
compute loss for weight  0.949777  0.949787 result 1.38803
 training batch 16 mu var0-0.272543
compute loss for weight  0.949792  0.949787 result 1.38802
 training batch 17 mu var0-0.272543
compute loss for weight  0.949782  0.949787 result 1.38803
   --dy = -0.314145 dy_ref = -0.314145
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      3.248     -0.4715 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.272543
compute loss for weight  1.00001  1 result 1.38806
 training batch 19 mu var0-0.272543
compute loss for weight  0.99999  1 result 1.38799
 training batch 20 mu var0-0.272543
compute loss for weight  1.00001  1 result 1.38804
 training batch 21 mu var0-0.272543
compute loss for weight  0.999995  1 result 1.38801
   --dy = 3.24759 dy_ref = 3.24759
 training batch 22 mu var0-0.272543
compute loss for weight  1.00001  1 result 1.38802
 training batch 23 mu var0-0.272543
compute loss for weight  0.99999  1 result 1.38803
 training batch 24 mu var0-0.272543
compute loss for weight  1.00001  1 result 1.38802
 training batch 25 mu var0-0.272543
compute loss for weight  0.999995  1 result 1.38803
   --dy = -0.471543 dy_ref = -0.471543
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -8.327e-17  -3.123e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.272543
compute loss for weight  1e-05  0 result 1.38803
 training batch 27 mu var0-0.272543
compute loss for weight  -1e-05  0 result 1.38803
 training batch 28 mu var0-0.272543
compute loss for weight  5e-06  0 result 1.38803
 training batch 29 mu var0-0.272543
compute loss for weight  -5e-06  0 result 1.38803
   --dy = 5.55112e-11 dy_ref = -8.32667e-17
 training batch 30 mu var0-0.272543
compute loss for weight  1e-05  0 result 1.38803
 training batch 31 mu var0-0.272543
compute loss for weight  -1e-05  0 result 1.38803
 training batch 32 mu var0-0.272543
compute loss for weight  5e-06  0 result 1.38803
 training batch 33 mu var0-0.272543
compute loss for weight  -5e-06  0 result 1.38803
   --dy = 5.18104e-11 dy_ref = -3.1225e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      2.239      0.9555 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |       1.45     -0.4935 

 training batch 34 mu var0-0.272543
compute loss for weight  1.45023  1.45022 result 1.38805
 training batch 35 mu var0-0.272543
compute loss for weight  1.45021  1.45022 result 1.388
 training batch 36 mu var0-0.272543
compute loss for weight  1.45023  1.45022 result 1.38804
 training batch 37 mu var0-0.272543
compute loss for weight  1.45022  1.45022 result 1.38801
   --dy = 2.23938 dy_ref = 2.23938
 training batch 38 mu var0-0.272543
compute loss for weight  -0.493509  -0.493519 result 1.38804
 training batch 39 mu var0-0.272543
compute loss for weight  -0.493529  -0.493519 result 1.38802
 training batch 40 mu var0-0.272543
compute loss for weight  -0.493514  -0.493519 result 1.38803
 training batch 41 mu var0-0.272543
compute loss for weight  -0.493524  -0.493519 result 1.38802
   --dy = 0.955469 dy_ref = 0.955469
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m2.2266e-09[NON-XML-CHAR-0x1B][39m