Execution Time0.13s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-centos7-clang100-opt-master (olhswep09.cern.ch) on 2019-11-13 23:47:29
Repository revision: 32b17abcda23e44b64218a42d0ca69cb30cda7e0

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 var0-0.152919
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.2863      0.9252 
   1 |    -0.9357     0.07693 
   2 |     0.7323      0.4156 
   3 |      1.162      -2.242 
   4 |     0.1629      0.2744 
   5 |    -0.3825       1.263 
   6 |     0.6199     -0.5992 
   7 |     -1.762       1.742 
   8 |     -1.083       2.475 
   9 |     0.2433     -0.5096 

output BN 
output DL feature 0 mean -0.152919	output DL std 0.913171
output DL feature 1 mean 0.382122	output DL std 1.33531
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -0.154      0.4287 
   1 |    -0.9035     -0.2409 
   2 |      1.022     0.02643 
   3 |      1.518      -2.071 
   4 |     0.3645      -0.085 
   5 |     -0.265       0.695 
   6 |     0.8921     -0.7746 
   7 |     -1.858       1.074 
   8 |     -1.073       1.652 
   9 |     0.4573     -0.7039 

output BN feature 0 mean 6.10623e-17	output BN std 1.05402
output BN feature 1 mean 1.11022e-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 |      0.418     -0.1117       1.034      0.2384 
   1 |     0.1493     -0.3386      0.5666     0.06521 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    -0.9186     -0.9893      0.2583     0.02655 
   1 |     0.8621      0.9904      0.6353     0.06925 

 training batch 2 mu var0-0.152916
compute loss for weight  -0.918552  -0.918562 result 0.500056
 training batch 3 mu var0-0.152919
compute loss for weight  -0.918572  -0.918562 result 0.500048
 training batch 4 mu var0-0.152918
compute loss for weight  -0.918557  -0.918562 result 0.500054
 training batch 5 mu var0-0.152919
compute loss for weight  -0.918567  -0.918562 result 0.50005
   --dy = 0.418048 dy_ref = 0.418048
 training batch 6 mu var0-0.152919
compute loss for weight  -0.989307  -0.989317 result 0.500051
 training batch 7 mu var0-0.152919
compute loss for weight  -0.989327  -0.989317 result 0.500053
 training batch 8 mu var0-0.152919
compute loss for weight  -0.989312  -0.989317 result 0.500051
 training batch 9 mu var0-0.152919
compute loss for weight  -0.989322  -0.989317 result 0.500053
   --dy = -0.111672 dy_ref = -0.111672
 training batch 10 mu var0-0.152918
compute loss for weight  0.258314  0.258304 result 0.500062
 training batch 11 mu var0-0.152919
compute loss for weight  0.258294  0.258304 result 0.500042
 training batch 12 mu var0-0.152919
compute loss for weight  0.258309  0.258304 result 0.500057
 training batch 13 mu var0-0.152919
compute loss for weight  0.258299  0.258304 result 0.500047
   --dy = 1.03431 dy_ref = 1.03431
 training batch 14 mu var0-0.152919
compute loss for weight  0.0265561  0.0265461 result 0.500054
 training batch 15 mu var0-0.152919
compute loss for weight  0.0265361  0.0265461 result 0.50005
 training batch 16 mu var0-0.152919
compute loss for weight  0.0265511  0.0265461 result 0.500053
 training batch 17 mu var0-0.152919
compute loss for weight  0.0265411  0.0265461 result 0.500051
   --dy = 0.238379 dy_ref = 0.238379
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.2308       1.231 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.152919
compute loss for weight  1.00001  1 result 0.50005
 training batch 19 mu var0-0.152919
compute loss for weight  0.99999  1 result 0.500054
 training batch 20 mu var0-0.152919
compute loss for weight  1.00001  1 result 0.500051
 training batch 21 mu var0-0.152919
compute loss for weight  0.999995  1 result 0.500053
   --dy = -0.230754 dy_ref = -0.230754
 training batch 22 mu var0-0.152919
compute loss for weight  1.00001  1 result 0.500064
 training batch 23 mu var0-0.152919
compute loss for weight  0.99999  1 result 0.50004
 training batch 24 mu var0-0.152919
compute loss for weight  1.00001  1 result 0.500058
 training batch 25 mu var0-0.152919
compute loss for weight  0.999995  1 result 0.500046
   --dy = 1.23086 dy_ref = 1.23086
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  2.776e-17   8.327e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.152919
compute loss for weight  1e-05  0 result 0.500052
 training batch 27 mu var0-0.152919
compute loss for weight  -1e-05  0 result 0.500052
 training batch 28 mu var0-0.152919
compute loss for weight  5e-06  0 result 0.500052
 training batch 29 mu var0-0.152919
compute loss for weight  -5e-06  0 result 0.500052
   --dy = -1.29526e-11 dy_ref = 2.77556e-17
 training batch 30 mu var0-0.152919
compute loss for weight  1e-05  0 result 0.500052
 training batch 31 mu var0-0.152919
compute loss for weight  -1e-05  0 result 0.500052
 training batch 32 mu var0-0.152919
compute loss for weight  5e-06  0 result 0.500052
 training batch 33 mu var0-0.152919
compute loss for weight  -5e-06  0 result 0.500052
   --dy = 1.4803e-11 dy_ref = 8.32667e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.3054      -1.079 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.7557      -1.141 

 training batch 34 mu var0-0.152919
compute loss for weight  -0.755642  -0.755652 result 0.500055
 training batch 35 mu var0-0.152919
compute loss for weight  -0.755662  -0.755652 result 0.500049
 training batch 36 mu var0-0.152919
compute loss for weight  -0.755647  -0.755652 result 0.500053
 training batch 37 mu var0-0.152919
compute loss for weight  -0.755657  -0.755652 result 0.50005
   --dy = 0.305371 dy_ref = 0.305371
 training batch 38 mu var0-0.152919
compute loss for weight  -1.14096  -1.14097 result 0.500041
 training batch 39 mu var0-0.152919
compute loss for weight  -1.14098  -1.14097 result 0.500063
 training batch 40 mu var0-0.152919
compute loss for weight  -1.14097  -1.14097 result 0.500047
 training batch 41 mu var0-0.152919
compute loss for weight  -1.14098  -1.14097 result 0.500057
   --dy = -1.07878 dy_ref = -1.07878
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m5.47589e-11[NON-XML-CHAR-0x1B][39m