Execution Time0.10s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-ubuntu19-gcc9 (root-ubuntu-1910-1) on 2019-11-15 00:48:29
Repository revision: 14de58de35eff907054671888ccc2de0f7f27e77

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 var00.535059
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.8817     -0.5657 
   1 |     0.8183     -0.2504 
   2 |     0.4414     -0.3811 
   3 |     0.7028       0.491 
   4 |     0.8596     -0.3597 
   5 |      2.036       -1.03 
   6 |     -1.034      0.7237 
   7 |   -0.01585     -0.7185 
   8 |     0.7106     -0.9445 
   9 |   -0.04982      0.1884 

output BN 
output DL feature 0 mean 0.535059	output DL std 0.798218
output DL feature 1 mean -0.284635	output DL std 0.588399
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.4577     -0.5033 
   1 |      0.374     0.06134 
   2 |    -0.1236     -0.1729 
   3 |     0.2215       1.389 
   4 |     0.4285     -0.1345 
   5 |      1.982      -1.334 
   6 |     -2.072       1.806 
   7 |    -0.7274     -0.7772 
   8 |     0.2318      -1.182 
   9 |    -0.7723      0.8473 

output BN feature 0 mean 0	output BN std 1.054
output BN feature 1 mean 1.11022e-17	output BN std 1.05392
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    0.04009      0.1557      0.1263     0.05967 
   1 |     0.4667      0.5104      0.7002     0.04711 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.8997   -0.004212    -0.04338     -0.5004 
   1 |    -0.5509     -0.3781     -0.2212      0.2278 

 training batch 2 mu var00.535061
compute loss for weight  0.899691  0.899681 result 0.394242
 training batch 3 mu var00.535059
compute loss for weight  0.899671  0.899681 result 0.394241
 training batch 4 mu var00.535059
compute loss for weight  0.899686  0.899681 result 0.394242
 training batch 5 mu var00.535059
compute loss for weight  0.899676  0.899681 result 0.394241
   --dy = 0.0400896 dy_ref = 0.0400896
 training batch 6 mu var00.535058
compute loss for weight  -0.00420205  -0.00421205 result 0.394243
 training batch 7 mu var00.535059
compute loss for weight  -0.00422205  -0.00421205 result 0.39424
 training batch 8 mu var00.535058
compute loss for weight  -0.00420705  -0.00421205 result 0.394242
 training batch 9 mu var00.535059
compute loss for weight  -0.00421705  -0.00421205 result 0.394241
   --dy = 0.155694 dy_ref = 0.155694
 training batch 10 mu var00.535059
compute loss for weight  -0.0433737  -0.0433837 result 0.394243
 training batch 11 mu var00.535059
compute loss for weight  -0.0433937  -0.0433837 result 0.39424
 training batch 12 mu var00.535059
compute loss for weight  -0.0433787  -0.0433837 result 0.394242
 training batch 13 mu var00.535059
compute loss for weight  -0.0433887  -0.0433837 result 0.394241
   --dy = 0.126292 dy_ref = 0.126292
 training batch 14 mu var00.535059
compute loss for weight  -0.50039  -0.5004 result 0.394242
 training batch 15 mu var00.535059
compute loss for weight  -0.50041  -0.5004 result 0.394241
 training batch 16 mu var00.535059
compute loss for weight  -0.500395  -0.5004 result 0.394242
 training batch 17 mu var00.535059
compute loss for weight  -0.500405  -0.5004 result 0.394241
   --dy = 0.0596656 dy_ref = 0.0596656
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.4383      0.3502 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.535059
compute loss for weight  1.00001  1 result 0.394246
 training batch 19 mu var00.535059
compute loss for weight  0.99999  1 result 0.394237
 training batch 20 mu var00.535059
compute loss for weight  1.00001  1 result 0.394244
 training batch 21 mu var00.535059
compute loss for weight  0.999995  1 result 0.39424
   --dy = 0.438301 dy_ref = 0.438301
 training batch 22 mu var00.535059
compute loss for weight  1.00001  1 result 0.394245
 training batch 23 mu var00.535059
compute loss for weight  0.99999  1 result 0.394238
 training batch 24 mu var00.535059
compute loss for weight  1.00001  1 result 0.394243
 training batch 25 mu var00.535059
compute loss for weight  0.999995  1 result 0.39424
   --dy = 0.350182 dy_ref = 0.350182
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  2.082e-17  -1.388e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.535059
compute loss for weight  1e-05  0 result 0.394242
 training batch 27 mu var00.535059
compute loss for weight  -1e-05  0 result 0.394242
 training batch 28 mu var00.535059
compute loss for weight  5e-06  0 result 0.394242
 training batch 29 mu var00.535059
compute loss for weight  -5e-06  0 result 0.394242
   --dy = 6.4763e-12 dy_ref = 2.08167e-17
 training batch 30 mu var00.535059
compute loss for weight  1e-05  0 result 0.394242
 training batch 31 mu var00.535059
compute loss for weight  -1e-05  0 result 0.394242
 training batch 32 mu var00.535059
compute loss for weight  5e-06  0 result 0.394242
 training batch 33 mu var00.535059
compute loss for weight  -5e-06  0 result 0.394242
   --dy = 9.25186e-13 dy_ref = -1.38778e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.167       1.125 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.3755      0.3114 

 training batch 34 mu var00.535059
compute loss for weight  -0.375495  -0.375505 result 0.39423
 training batch 35 mu var00.535059
compute loss for weight  -0.375515  -0.375505 result 0.394253
 training batch 36 mu var00.535059
compute loss for weight  -0.3755  -0.375505 result 0.394236
 training batch 37 mu var00.535059
compute loss for weight  -0.37551  -0.375505 result 0.394248
   --dy = -1.16723 dy_ref = -1.16723
 training batch 38 mu var00.535059
compute loss for weight  0.311379  0.311369 result 0.394253
 training batch 39 mu var00.535059
compute loss for weight  0.311359  0.311369 result 0.39423
 training batch 40 mu var00.535059
compute loss for weight  0.311374  0.311369 result 0.394247
 training batch 41 mu var00.535059
compute loss for weight  0.311364  0.311369 result 0.394236
   --dy = 1.12465 dy_ref = 1.12465
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m2.53023e-10[NON-XML-CHAR-0x1B][39m