Execution Time1.05s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: PR-4625-x86_64-centos7-gcc48-opt (olhswep22.cern.ch) on 2019-11-15 11:07:26

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 var0-0.211413
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.1631     0.04972 
   1 |    -0.1208       1.441 
   2 |     -1.242      -1.166 
   3 |     -2.748       2.489 
   4 |    -0.8295       0.438 
   5 |     -1.073      0.9066 
   6 |      1.589      0.7051 
   7 |      1.053      -2.093 
   8 |      1.751     -0.9652 
   9 |    -0.3309      0.4458 

output BN 
output DL feature 0 mean -0.211413	output DL std 1.3878
output DL feature 1 mean 0.22513	output DL std 1.33956
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    0.03668      -0.138 
   1 |    0.06879      0.9564 
   2 |    -0.7826      -1.094 
   3 |     -1.926       1.782 
   4 |    -0.4695      0.1675 
   5 |    -0.6543      0.5363 
   6 |      1.368      0.3777 
   7 |     0.9601      -1.824 
   8 |       1.49     -0.9366 
   9 |   -0.09076      0.1736 

output BN feature 0 mean 1.11022e-17	output BN std 1.05406
output BN feature 1 mean -6.38378e-17	output BN std 1.05406
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |   -0.01654     0.06698     0.07456    -0.05278 
   1 |    -0.1333        1.11      0.8881      0.1125 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.1221      0.7723      0.1068        1.09 
   1 |     0.4848     -0.9773      -1.098     0.04443 

 training batch 2 mu var0-0.21141
compute loss for weight  0.122071  0.122061 result 1.10396
 training batch 3 mu var0-0.211413
compute loss for weight  0.122051  0.122061 result 1.10396
 training batch 4 mu var0-0.211412
compute loss for weight  0.122066  0.122061 result 1.10396
 training batch 5 mu var0-0.211413
compute loss for weight  0.122056  0.122061 result 1.10396
   --dy = -0.0165445 dy_ref = -0.0165445
 training batch 6 mu var0-0.211413
compute loss for weight  0.772272  0.772262 result 1.10396
 training batch 7 mu var0-0.211413
compute loss for weight  0.772252  0.772262 result 1.10396
 training batch 8 mu var0-0.211413
compute loss for weight  0.772267  0.772262 result 1.10396
 training batch 9 mu var0-0.211413
compute loss for weight  0.772257  0.772262 result 1.10396
   --dy = 0.066984 dy_ref = 0.066984
 training batch 10 mu var0-0.211412
compute loss for weight  0.106784  0.106774 result 1.10396
 training batch 11 mu var0-0.211413
compute loss for weight  0.106764  0.106774 result 1.10396
 training batch 12 mu var0-0.211412
compute loss for weight  0.106779  0.106774 result 1.10396
 training batch 13 mu var0-0.211413
compute loss for weight  0.106769  0.106774 result 1.10396
   --dy = 0.0745627 dy_ref = 0.0745627
 training batch 14 mu var0-0.211413
compute loss for weight  1.09043  1.09042 result 1.10396
 training batch 15 mu var0-0.211413
compute loss for weight  1.09041  1.09042 result 1.10396
 training batch 16 mu var0-0.211413
compute loss for weight  1.09042  1.09042 result 1.10396
 training batch 17 mu var0-0.211413
compute loss for weight  1.09041  1.09042 result 1.10396
   --dy = -0.0527808 dy_ref = -0.0527808
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      2.046      0.1623 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.211413
compute loss for weight  1.00001  1 result 1.10398
 training batch 19 mu var0-0.211413
compute loss for weight  0.99999  1 result 1.10394
 training batch 20 mu var0-0.211413
compute loss for weight  1.00001  1 result 1.10397
 training batch 21 mu var0-0.211413
compute loss for weight  0.999995  1 result 1.10395
   --dy = 2.04559 dy_ref = 2.04559
 training batch 22 mu var0-0.211413
compute loss for weight  1.00001  1 result 1.10396
 training batch 23 mu var0-0.211413
compute loss for weight  0.99999  1 result 1.10396
 training batch 24 mu var0-0.211413
compute loss for weight  1.00001  1 result 1.10396
 training batch 25 mu var0-0.211413
compute loss for weight  0.999995  1 result 1.10396
   --dy = 0.162323 dy_ref = 0.162323
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -5.551e-17  -6.939e-18 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.211413
compute loss for weight  1e-05  0 result 1.10396
 training batch 27 mu var0-0.211413
compute loss for weight  -1e-05  0 result 1.10396
 training batch 28 mu var0-0.211413
compute loss for weight  5e-06  0 result 1.10396
 training batch 29 mu var0-0.211413
compute loss for weight  -5e-06  0 result 1.10396
   --dy = -6.66134e-11 dy_ref = -5.55112e-17
 training batch 30 mu var0-0.211413
compute loss for weight  1e-05  0 result 1.10396
 training batch 31 mu var0-0.211413
compute loss for weight  -1e-05  0 result 1.10396
 training batch 32 mu var0-0.211413
compute loss for weight  5e-06  0 result 1.10396
 training batch 33 mu var0-0.211413
compute loss for weight  -5e-06  0 result 1.10396
   --dy = 3.33067e-11 dy_ref = -6.93889e-18
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      2.092      -1.336 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      0.978     -0.1215 

 training batch 34 mu var0-0.211413
compute loss for weight  0.977988  0.977978 result 1.10398
 training batch 35 mu var0-0.211413
compute loss for weight  0.977968  0.977978 result 1.10394
 training batch 36 mu var0-0.211413
compute loss for weight  0.977983  0.977978 result 1.10397
 training batch 37 mu var0-0.211413
compute loss for weight  0.977973  0.977978 result 1.10395
   --dy = 2.09166 dy_ref = 2.09166
 training batch 38 mu var0-0.211413
compute loss for weight  -0.121447  -0.121457 result 1.10394
 training batch 39 mu var0-0.211413
compute loss for weight  -0.121467  -0.121457 result 1.10397
 training batch 40 mu var0-0.211413
compute loss for weight  -0.121452  -0.121457 result 1.10395
 training batch 41 mu var0-0.211413
compute loss for weight  -0.121462  -0.121457 result 1.10396
   --dy = -1.33647 dy_ref = -1.33647
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m4.06135e-09[NON-XML-CHAR-0x1B][39m