Execution Time0.17s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-centos7-clang100-opt-no-rt-cxxmodules (olsnba08.cern.ch) on 2019-11-15 13:39:49
Repository revision: ddaf537cc1431ddcb8dc5b394576c131f26a6e1a

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 var00.185788
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.4714      -1.391 
   1 |     0.7008      0.3906 
   2 |     -1.844      -1.592 
   3 |     -2.336    -0.03929 
   4 |    -0.2887      -1.326 
   5 |     0.3733      -2.716 
   6 |      2.511      0.1552 
   7 |    -0.2133       1.206 
   8 |      2.773      -1.834 
   9 |     -0.289      0.1915 

output BN 
output DL feature 0 mean 0.185788	output DL std 1.61812
output DL feature 1 mean -0.695419	output DL std 1.23818
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1861     -0.5922 
   1 |     0.3355      0.9245 
   2 |     -1.322     -0.7635 
   3 |     -1.642      0.5586 
   4 |    -0.3091     -0.5365 
   5 |     0.1221       -1.72 
   6 |      1.514      0.7241 
   7 |      -0.26       1.619 
   8 |      1.685      -0.969 
   9 |    -0.3093       0.755 

output BN feature 0 mean 4.996e-17	output BN std 1.05407
output BN feature 1 mean 3.33067e-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 |     -2.184       1.956      -2.679      0.4302 
   1 |        3.8     -0.9219       3.768       4.259 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.7977      0.3042     -0.1901       1.483 
   1 |     -0.444      0.5776     -0.8862      0.1205 

 training batch 2 mu var00.185791
compute loss for weight  0.797757  0.797747 result 5.674
 training batch 3 mu var00.185788
compute loss for weight  0.797737  0.797747 result 5.67404
 training batch 4 mu var00.185789
compute loss for weight  0.797752  0.797747 result 5.67401
 training batch 5 mu var00.185788
compute loss for weight  0.797742  0.797747 result 5.67403
   --dy = -2.18408 dy_ref = -2.18408
 training batch 6 mu var00.185788
compute loss for weight  0.304224  0.304214 result 5.67404
 training batch 7 mu var00.185788
compute loss for weight  0.304204  0.304214 result 5.674
 training batch 8 mu var00.185788
compute loss for weight  0.304219  0.304214 result 5.67403
 training batch 9 mu var00.185788
compute loss for weight  0.304209  0.304214 result 5.67401
   --dy = 1.9563 dy_ref = 1.9563
 training batch 10 mu var00.185789
compute loss for weight  -0.190086  -0.190096 result 5.674
 training batch 11 mu var00.185788
compute loss for weight  -0.190106  -0.190096 result 5.67405
 training batch 12 mu var00.185788
compute loss for weight  -0.190091  -0.190096 result 5.67401
 training batch 13 mu var00.185788
compute loss for weight  -0.190101  -0.190096 result 5.67404
   --dy = -2.67928 dy_ref = -2.67928
 training batch 14 mu var00.185788
compute loss for weight  1.48314  1.48313 result 5.67403
 training batch 15 mu var00.185788
compute loss for weight  1.48312  1.48313 result 5.67402
 training batch 16 mu var00.185788
compute loss for weight  1.48313  1.48313 result 5.67402
 training batch 17 mu var00.185788
compute loss for weight  1.48312  1.48313 result 5.67402
   --dy = 0.43021 dy_ref = 0.43021
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      3.862       7.486 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.185788
compute loss for weight  1.00001  1 result 5.67406
 training batch 19 mu var00.185788
compute loss for weight  0.99999  1 result 5.67398
 training batch 20 mu var00.185788
compute loss for weight  1.00001  1 result 5.67404
 training batch 21 mu var00.185788
compute loss for weight  0.999995  1 result 5.674
   --dy = 3.86182 dy_ref = 3.86182
 training batch 22 mu var00.185788
compute loss for weight  1.00001  1 result 5.6741
 training batch 23 mu var00.185788
compute loss for weight  0.99999  1 result 5.67395
 training batch 24 mu var00.185788
compute loss for weight  1.00001  1 result 5.67406
 training batch 25 mu var00.185788
compute loss for weight  0.999995  1 result 5.67398
   --dy = 7.48622 dy_ref = 7.48622
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  5.551e-17    1.11e-16 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.185788
compute loss for weight  1e-05  0 result 5.67402
 training batch 27 mu var00.185788
compute loss for weight  -1e-05  0 result 5.67402
 training batch 28 mu var00.185788
compute loss for weight  5e-06  0 result 5.67402
 training batch 29 mu var00.185788
compute loss for weight  -5e-06  0 result 5.67402
   --dy = -1.4803e-11 dy_ref = 5.55112e-17
 training batch 30 mu var00.185788
compute loss for weight  1e-05  0 result 5.67402
 training batch 31 mu var00.185788
compute loss for weight  -1e-05  0 result 5.67402
 training batch 32 mu var00.185788
compute loss for weight  5e-06  0 result 5.67402
 training batch 33 mu var00.185788
compute loss for weight  -5e-06  0 result 5.67402
   --dy = -1.4803e-11 dy_ref = 1.11022e-16
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      2.596       3.732 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.487       2.006 

 training batch 34 mu var00.185788
compute loss for weight  1.48738  1.48737 result 5.67405
 training batch 35 mu var00.185788
compute loss for weight  1.48736  1.48737 result 5.674
 training batch 36 mu var00.185788
compute loss for weight  1.48738  1.48737 result 5.67403
 training batch 37 mu var00.185788
compute loss for weight  1.48737  1.48737 result 5.67401
   --dy = 2.5964 dy_ref = 2.5964
 training batch 38 mu var00.185788
compute loss for weight  2.00617  2.00616 result 5.67406
 training batch 39 mu var00.185788
compute loss for weight  2.00615  2.00616 result 5.67398
 training batch 40 mu var00.185788
compute loss for weight  2.00617  2.00616 result 5.67404
 training batch 41 mu var00.185788
compute loss for weight  2.00616  2.00616 result 5.674
   --dy = 3.73162 dy_ref = 3.73162
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m3.59845e-10[NON-XML-CHAR-0x1B][39m