Execution Time0.31s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-mac1014-clang100 (macphsft17.dyndns.cern.ch) on 2019-11-14 02:59:38

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 var0-0.567513
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.8466    0.002859 
   1 |     0.2338      0.5626 
   2 |     -2.383      -1.948 
   3 |     -3.322      -1.007 
   4 |     -1.645     -0.3659 
   5 |     -2.535     -0.2289 
   6 |      1.843       2.908 
   7 |      2.047      -1.392 
   8 |      1.245       1.413 
   9 |    -0.3132    0.003569 

output BN 
output DL feature 0 mean -0.567513	output DL std 1.90419
output DL feature 1 mean -0.00514637	output DL std 1.40263
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.1545    0.006016 
   1 |     0.4436      0.4266 
   2 |     -1.005       -1.46 
   3 |     -1.525     -0.7526 
   4 |    -0.5965     -0.2711 
   5 |     -1.089     -0.1681 
   6 |      1.335       2.189 
   7 |      1.447      -1.042 
   8 |      1.004       1.066 
   9 |     0.1408    0.006549 

output BN feature 0 mean -9.4369e-17	output BN std 1.05408
output BN feature 1 mean -4.51895e-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 |     0.1186      -0.255    -0.01985      0.2414 
   1 |    -0.2743      0.6065     0.04179     -0.5139 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    0.06862       1.314     -0.3891       1.322 
   1 |     0.2662     -0.3567     -0.4789       1.452 

 training batch 2 mu var0-0.56751
compute loss for weight  0.0686272  0.0686172 result 2.59964
 training batch 3 mu var0-0.567513
compute loss for weight  0.0686072  0.0686172 result 2.59963
 training batch 4 mu var0-0.567513
compute loss for weight  0.0686222  0.0686172 result 2.59964
 training batch 5 mu var0-0.567513
compute loss for weight  0.0686122  0.0686172 result 2.59963
   --dy = 0.118594 dy_ref = 0.118594
 training batch 6 mu var0-0.567514
compute loss for weight  1.31384  1.31383 result 2.59963
 training batch 7 mu var0-0.567513
compute loss for weight  1.31382  1.31383 result 2.59964
 training batch 8 mu var0-0.567513
compute loss for weight  1.31383  1.31383 result 2.59963
 training batch 9 mu var0-0.567513
compute loss for weight  1.31382  1.31383 result 2.59964
   --dy = -0.25499 dy_ref = -0.25499
 training batch 10 mu var0-0.567513
compute loss for weight  -0.389114  -0.389124 result 2.59963
 training batch 11 mu var0-0.567513
compute loss for weight  -0.389134  -0.389124 result 2.59964
 training batch 12 mu var0-0.567513
compute loss for weight  -0.389119  -0.389124 result 2.59964
 training batch 13 mu var0-0.567513
compute loss for weight  -0.389129  -0.389124 result 2.59964
   --dy = -0.0198511 dy_ref = -0.0198511
 training batch 14 mu var0-0.567513
compute loss for weight  1.32195  1.32194 result 2.59964
 training batch 15 mu var0-0.567513
compute loss for weight  1.32193  1.32194 result 2.59963
 training batch 16 mu var0-0.567513
compute loss for weight  1.32195  1.32194 result 2.59964
 training batch 17 mu var0-0.567513
compute loss for weight  1.32194  1.32194 result 2.59963
   --dy = 0.241444 dy_ref = 0.241444
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.7779       4.421 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.567513
compute loss for weight  1.00001  1 result 2.59964
 training batch 19 mu var0-0.567513
compute loss for weight  0.99999  1 result 2.59963
 training batch 20 mu var0-0.567513
compute loss for weight  1.00001  1 result 2.59964
 training batch 21 mu var0-0.567513
compute loss for weight  0.999995  1 result 2.59963
   --dy = 0.777858 dy_ref = 0.777858
 training batch 22 mu var0-0.567513
compute loss for weight  1.00001  1 result 2.59968
 training batch 23 mu var0-0.567513
compute loss for weight  0.99999  1 result 2.59959
 training batch 24 mu var0-0.567513
compute loss for weight  1.00001  1 result 2.59966
 training batch 25 mu var0-0.567513
compute loss for weight  0.999995  1 result 2.59961
   --dy = 4.42141 dy_ref = 4.42141
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -3.643e-17  -1.908e-16 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.567513
compute loss for weight  1e-05  0 result 2.59964
 training batch 27 mu var0-0.567513
compute loss for weight  -1e-05  0 result 2.59964
 training batch 28 mu var0-0.567513
compute loss for weight  5e-06  0 result 2.59964
 training batch 29 mu var0-0.567513
compute loss for weight  -5e-06  0 result 2.59964
   --dy = -7.40149e-12 dy_ref = -3.64292e-17
 training batch 30 mu var0-0.567513
compute loss for weight  1e-05  0 result 2.59964
 training batch 31 mu var0-0.567513
compute loss for weight  -1e-05  0 result 2.59964
 training batch 32 mu var0-0.567513
compute loss for weight  5e-06  0 result 2.59964
 training batch 33 mu var0-0.567513
compute loss for weight  -5e-06  0 result 2.59964
   --dy = -5.18104e-11 dy_ref = -1.9082e-16
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -2.258      -3.174 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.3445      -1.393 

 training batch 34 mu var0-0.567513
compute loss for weight  -0.344474  -0.344484 result 2.59961
 training batch 35 mu var0-0.567513
compute loss for weight  -0.344494  -0.344484 result 2.59966
 training batch 36 mu var0-0.567513
compute loss for weight  -0.344479  -0.344484 result 2.59962
 training batch 37 mu var0-0.567513
compute loss for weight  -0.344489  -0.344484 result 2.59965
   --dy = -2.25804 dy_ref = -2.25804
 training batch 38 mu var0-0.567513
compute loss for weight  -1.39303  -1.39304 result 2.5996
 training batch 39 mu var0-0.567513
compute loss for weight  -1.39305  -1.39304 result 2.59967
 training batch 40 mu var0-0.567513
compute loss for weight  -1.39303  -1.39304 result 2.59962
 training batch 41 mu var0-0.567513
compute loss for weight  -1.39304  -1.39304 result 2.59965
   --dy = -3.17394 dy_ref = -3.17394
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m1.32176e-09[NON-XML-CHAR-0x1B][39m