Execution Time0.08s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-fedora29-gcc8-opt-exp-pyroot (root-fedora29-3.cern.ch) on 2019-11-13 05:17:19

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 var0-0.0449214
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1163      0.4571 
   1 |    -0.8932      -1.489 
   2 |     0.2517      0.8297 
   3 |     -1.166     -0.7358 
   4 |   -0.06521      0.4266 
   5 |    -0.1959      0.5057 
   6 |      1.255       2.262 
   7 |    -0.6158      -2.424 
   8 |      1.009       1.391 
   9 |    -0.1451    -0.07569 

output BN 
output DL feature 0 mean -0.0449214	output DL std 0.763841
output DL feature 1 mean 0.114642	output DL std 1.37012
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2225      0.2635 
   1 |     -1.171      -1.234 
   2 |     0.4093      0.5501 
   3 |     -1.547     -0.6542 
   4 |     -0.028        0.24 
   5 |    -0.2083      0.3008 
   6 |      1.794       1.652 
   7 |    -0.7877      -1.953 
   8 |      1.455      0.9818 
   9 |    -0.1382     -0.1464 

output BN feature 0 mean 4.44089e-17	output BN std 1.05399
output BN feature 1 mean 2.77556e-18	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 |     0.6075      -2.939      0.0615      -0.679 
   1 |    -0.9257       3.176       -1.32      -1.071 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    -0.4115     -0.2231      0.5754      0.6468 
   1 |    -0.7449      -1.102      0.9607      0.9682 

 training batch 2 mu var0-0.0449186
compute loss for weight  -0.411498  -0.411508 result 8.85214
 training batch 3 mu var0-0.0449214
compute loss for weight  -0.411518  -0.411508 result 8.85213
 training batch 4 mu var0-0.0449207
compute loss for weight  -0.411503  -0.411508 result 8.85214
 training batch 5 mu var0-0.0449214
compute loss for weight  -0.411513  -0.411508 result 8.85213
   --dy = 0.607497 dy_ref = 0.607497
 training batch 6 mu var0-0.0449219
compute loss for weight  -0.223054  -0.223064 result 8.8521
 training batch 7 mu var0-0.0449214
compute loss for weight  -0.223074  -0.223064 result 8.85216
 training batch 8 mu var0-0.0449216
compute loss for weight  -0.223059  -0.223064 result 8.85212
 training batch 9 mu var0-0.0449214
compute loss for weight  -0.223069  -0.223064 result 8.85215
   --dy = -2.93876 dy_ref = -2.93876
 training batch 10 mu var0-0.0449211
compute loss for weight  0.57542  0.57541 result 8.85213
 training batch 11 mu var0-0.0449214
compute loss for weight  0.5754  0.57541 result 8.85213
 training batch 12 mu var0-0.0449213
compute loss for weight  0.575415  0.57541 result 8.85213
 training batch 13 mu var0-0.0449214
compute loss for weight  0.575405  0.57541 result 8.85213
   --dy = 0.0615048 dy_ref = 0.0615048
 training batch 14 mu var0-0.0449215
compute loss for weight  0.646854  0.646844 result 8.85213
 training batch 15 mu var0-0.0449214
compute loss for weight  0.646834  0.646844 result 8.85214
 training batch 16 mu var0-0.0449214
compute loss for weight  0.646849  0.646844 result 8.85213
 training batch 17 mu var0-0.0449214
compute loss for weight  0.646839  0.646844 result 8.85214
   --dy = -0.678964 dy_ref = -0.678964
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      9.179       8.526 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.0449214
compute loss for weight  1.00001  1 result 8.85222
 training batch 19 mu var0-0.0449214
compute loss for weight  0.99999  1 result 8.85204
 training batch 20 mu var0-0.0449214
compute loss for weight  1.00001  1 result 8.85218
 training batch 21 mu var0-0.0449214
compute loss for weight  0.999995  1 result 8.85209
   --dy = 9.17858 dy_ref = 9.17858
 training batch 22 mu var0-0.0449214
compute loss for weight  1.00001  1 result 8.85222
 training batch 23 mu var0-0.0449214
compute loss for weight  0.99999  1 result 8.85205
 training batch 24 mu var0-0.0449214
compute loss for weight  1.00001  1 result 8.85217
 training batch 25 mu var0-0.0449214
compute loss for weight  0.999995  1 result 8.85209
   --dy = 8.52568 dy_ref = 8.52568
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  2.776e-16   1.388e-16 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.0449214
compute loss for weight  1e-05  0 result 8.85213
 training batch 27 mu var0-0.0449214
compute loss for weight  -1e-05  0 result 8.85213
 training batch 28 mu var0-0.0449214
compute loss for weight  5e-06  0 result 8.85213
 training batch 29 mu var0-0.0449214
compute loss for weight  -5e-06  0 result 8.85213
   --dy = -2.96059e-10 dy_ref = 2.77556e-16
 training batch 30 mu var0-0.0449214
compute loss for weight  1e-05  0 result 8.85213
 training batch 31 mu var0-0.0449214
compute loss for weight  -1e-05  0 result 8.85213
 training batch 32 mu var0-0.0449214
compute loss for weight  5e-06  0 result 8.85213
 training batch 33 mu var0-0.0449214
compute loss for weight  -5e-06  0 result 8.85213
   --dy = 2.36848e-10 dy_ref = 1.38778e-16
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      5.756       5.727 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.595       1.489 

 training batch 34 mu var0-0.0449214
compute loss for weight  1.59473  1.59472 result 8.85219
 training batch 35 mu var0-0.0449214
compute loss for weight  1.59471  1.59472 result 8.85207
 training batch 36 mu var0-0.0449214
compute loss for weight  1.59473  1.59472 result 8.85216
 training batch 37 mu var0-0.0449214
compute loss for weight  1.59472  1.59472 result 8.8521
   --dy = 5.7556 dy_ref = 5.7556
 training batch 38 mu var0-0.0449214
compute loss for weight  1.48875  1.48874 result 8.85219
 training batch 39 mu var0-0.0449214
compute loss for weight  1.48873  1.48874 result 8.85207
 training batch 40 mu var0-0.0449214
compute loss for weight  1.48874  1.48874 result 8.85216
 training batch 41 mu var0-0.0449214
compute loss for weight  1.48873  1.48874 result 8.8521
   --dy = 5.72679 dy_ref = 5.72679
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m1.59854e-08[NON-XML-CHAR-0x1B][39m