Execution Time1.17s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: PR-4624-x86_64-centos7-gcc48-opt (olhswep22.cern.ch) on 2019-11-14 19:00:31

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 var0-0.816318
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.596      0.3922 
   1 |    -0.2169      0.3915 
   2 |      -1.27      0.2702 
   3 |     0.4418        2.14 
   4 |     -1.307      0.8228 
   5 |     -3.075       1.374 
   6 |      0.754      0.2911 
   7 |     0.1412      -2.218 
   8 |     -2.355     -0.5578 
   9 |     0.3191      0.3088 

output BN 
output DL feature 0 mean -0.816318	output DL std 1.29686
output DL feature 1 mean 0.321553	output DL std 1.15011
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.6335     0.06472 
   1 |     0.4872     0.06407 
   2 |    -0.3688    -0.04708 
   3 |      1.023       1.667 
   4 |    -0.3988      0.4594 
   5 |     -1.836      0.9649 
   6 |      1.276    -0.02787 
   7 |     0.7783      -2.328 
   8 |      -1.25     -0.8059 
   9 |     0.9228    -0.01167 

output BN feature 0 mean 3.33067e-17	output BN std 1.05406
output BN feature 1 mean 5.01335e-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 |     0.2307      -2.017     -0.5938      -0.695 
   1 |     -2.181       2.335      -1.629       0.683 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     -1.025    0.003538     -0.7193      0.2637 
   1 |    0.09755      -1.134       -0.26     -0.1679 

 training batch 2 mu var0-0.816315
compute loss for weight  -1.02509  -1.0251 result 2.64868
 training batch 3 mu var0-0.816318
compute loss for weight  -1.02511  -1.0251 result 2.64868
 training batch 4 mu var0-0.816317
compute loss for weight  -1.0251  -1.0251 result 2.64868
 training batch 5 mu var0-0.816318
compute loss for weight  -1.02511  -1.0251 result 2.64868
   --dy = 0.230741 dy_ref = 0.230741
 training batch 6 mu var0-0.816318
compute loss for weight  0.00354823  0.00353823 result 2.64866
 training batch 7 mu var0-0.816318
compute loss for weight  0.00352823  0.00353823 result 2.6487
 training batch 8 mu var0-0.816318
compute loss for weight  0.00354323  0.00353823 result 2.64867
 training batch 9 mu var0-0.816318
compute loss for weight  0.00353323  0.00353823 result 2.64869
   --dy = -2.01737 dy_ref = -2.01737
 training batch 10 mu var0-0.816318
compute loss for weight  -0.719265  -0.719275 result 2.64867
 training batch 11 mu var0-0.816318
compute loss for weight  -0.719285  -0.719275 result 2.64868
 training batch 12 mu var0-0.816318
compute loss for weight  -0.71927  -0.719275 result 2.64868
 training batch 13 mu var0-0.816318
compute loss for weight  -0.71928  -0.719275 result 2.64868
   --dy = -0.593809 dy_ref = -0.593809
 training batch 14 mu var0-0.816318
compute loss for weight  0.263758  0.263748 result 2.64867
 training batch 15 mu var0-0.816318
compute loss for weight  0.263738  0.263748 result 2.64869
 training batch 16 mu var0-0.816318
compute loss for weight  0.263753  0.263748 result 2.64867
 training batch 17 mu var0-0.816318
compute loss for weight  0.263743  0.263748 result 2.64868
   --dy = -0.694951 dy_ref = -0.694951
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      2.263       3.034 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.816318
compute loss for weight  1.00001  1 result 2.6487
 training batch 19 mu var0-0.816318
compute loss for weight  0.99999  1 result 2.64866
 training batch 20 mu var0-0.816318
compute loss for weight  1.00001  1 result 2.64869
 training batch 21 mu var0-0.816318
compute loss for weight  0.999995  1 result 2.64867
   --dy = 2.26294 dy_ref = 2.26294
 training batch 22 mu var0-0.816318
compute loss for weight  1.00001  1 result 2.64871
 training batch 23 mu var0-0.816318
compute loss for weight  0.99999  1 result 2.64865
 training batch 24 mu var0-0.816318
compute loss for weight  1.00001  1 result 2.64869
 training batch 25 mu var0-0.816318
compute loss for weight  0.999995  1 result 2.64866
   --dy = 3.03442 dy_ref = 3.03442
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  -1.11e-16    1.11e-16 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.816318
compute loss for weight  1e-05  0 result 2.64868
 training batch 27 mu var0-0.816318
compute loss for weight  -1e-05  0 result 2.64868
 training batch 28 mu var0-0.816318
compute loss for weight  5e-06  0 result 2.64868
 training batch 29 mu var0-0.816318
compute loss for weight  -5e-06  0 result 2.64868
   --dy = 3.70074e-11 dy_ref = -1.11022e-16
 training batch 30 mu var0-0.816318
compute loss for weight  1e-05  0 result 2.64868
 training batch 31 mu var0-0.816318
compute loss for weight  -1e-05  0 result 2.64868
 training batch 32 mu var0-0.816318
compute loss for weight  5e-06  0 result 2.64868
 training batch 33 mu var0-0.816318
compute loss for weight  -5e-06  0 result 2.64868
   --dy = 0 dy_ref = 1.11022e-16
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.991      -2.342 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.137      -1.295 

 training batch 34 mu var0-0.816318
compute loss for weight  -1.13684  -1.13685 result 2.64866
 training batch 35 mu var0-0.816318
compute loss for weight  -1.13686  -1.13685 result 2.6487
 training batch 36 mu var0-0.816318
compute loss for weight  -1.13685  -1.13685 result 2.64867
 training batch 37 mu var0-0.816318
compute loss for weight  -1.13686  -1.13685 result 2.64869
   --dy = -1.99053 dy_ref = -1.99053
 training batch 38 mu var0-0.816318
compute loss for weight  -1.29546  -1.29547 result 2.64865
 training batch 39 mu var0-0.816318
compute loss for weight  -1.29548  -1.29547 result 2.6487
 training batch 40 mu var0-0.816318
compute loss for weight  -1.29546  -1.29547 result 2.64867
 training batch 41 mu var0-0.816318
compute loss for weight  -1.29547  -1.29547 result 2.64869
   --dy = -2.34234 dy_ref = -2.34234
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m2.38849e-10[NON-XML-CHAR-0x1B][39m