Execution Time0.07s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-fedora31-gcc9 (root-fedora-31-1.cern.ch) on 2019-11-14 00:48:30

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 var0-0.442964
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.7436      0.8201 
   1 |    0.06537    -0.09866 
   2 |     -1.689     -0.8706 
   3 |     -1.475     -0.7184 
   4 |     -1.068      0.5328 
   5 |     -1.869       1.373 
   6 |      1.759       3.715 
   7 |     0.3681      -3.319 
   8 |     0.2635       2.488 
   9 |   -0.04038    -0.02439 

output BN 
output DL feature 0 mean -0.442964	output DL std 1.13325
output DL feature 1 mean 0.389761	output DL std 1.93243
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.2796      0.2347 
   1 |     0.4728     -0.2664 
   2 |     -1.159     -0.6875 
   3 |      -0.96     -0.6045 
   4 |     -0.581     0.07804 
   5 |     -1.327      0.5361 
   6 |      2.048       1.814 
   7 |     0.7544      -2.023 
   8 |     0.6571       1.145 
   9 |     0.3744     -0.2259 

output BN feature 0 mean -6.66134e-17	output BN std 1.05405
output BN feature 1 mean -2.77556e-18	output BN std 1.05408
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    0.04944    -0.07179     0.03764     0.05363 
   1 |    -0.1097      0.1575    -0.08205      -0.141 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    -0.2422      0.3664     -0.3997      0.9956 
   1 |     0.1599      -1.284      0.1959       1.697 

 training batch 2 mu var0-0.442961
compute loss for weight  -0.24214  -0.24215 result 0.437591
 training batch 3 mu var0-0.442964
compute loss for weight  -0.24216  -0.24215 result 0.43759
 training batch 4 mu var0-0.442963
compute loss for weight  -0.242145  -0.24215 result 0.43759
 training batch 5 mu var0-0.442964
compute loss for weight  -0.242155  -0.24215 result 0.43759
   --dy = 0.0494374 dy_ref = 0.0494374
 training batch 6 mu var0-0.442964
compute loss for weight  0.366361  0.366351 result 0.437589
 training batch 7 mu var0-0.442964
compute loss for weight  0.366341  0.366351 result 0.437591
 training batch 8 mu var0-0.442964
compute loss for weight  0.366356  0.366351 result 0.43759
 training batch 9 mu var0-0.442964
compute loss for weight  0.366346  0.366351 result 0.437591
   --dy = -0.0717943 dy_ref = -0.0717943
 training batch 10 mu var0-0.442964
compute loss for weight  -0.399728  -0.399738 result 0.437591
 training batch 11 mu var0-0.442964
compute loss for weight  -0.399748  -0.399738 result 0.43759
 training batch 12 mu var0-0.442964
compute loss for weight  -0.399733  -0.399738 result 0.43759
 training batch 13 mu var0-0.442964
compute loss for weight  -0.399743  -0.399738 result 0.43759
   --dy = 0.0376368 dy_ref = 0.0376368
 training batch 14 mu var0-0.442964
compute loss for weight  0.995562  0.995552 result 0.437591
 training batch 15 mu var0-0.442964
compute loss for weight  0.995542  0.995552 result 0.43759
 training batch 16 mu var0-0.442964
compute loss for weight  0.995557  0.995552 result 0.43759
 training batch 17 mu var0-0.442964
compute loss for weight  0.995547  0.995552 result 0.43759
   --dy = 0.0536267 dy_ref = 0.0536267
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.8102     0.06502 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.442964
compute loss for weight  1.00001  1 result 0.437598
 training batch 19 mu var0-0.442964
compute loss for weight  0.99999  1 result 0.437582
 training batch 20 mu var0-0.442964
compute loss for weight  1.00001  1 result 0.437594
 training batch 21 mu var0-0.442964
compute loss for weight  0.999995  1 result 0.437586
   --dy = 0.81016 dy_ref = 0.81016
 training batch 22 mu var0-0.442964
compute loss for weight  1.00001  1 result 0.437591
 training batch 23 mu var0-0.442964
compute loss for weight  0.99999  1 result 0.43759
 training batch 24 mu var0-0.442964
compute loss for weight  1.00001  1 result 0.437591
 training batch 25 mu var0-0.442964
compute loss for weight  0.999995  1 result 0.43759
   --dy = 0.0650207 dy_ref = 0.0650207
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  1.041e-17  -8.674e-19 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.442964
compute loss for weight  1e-05  0 result 0.43759
 training batch 27 mu var0-0.442964
compute loss for weight  -1e-05  0 result 0.43759
 training batch 28 mu var0-0.442964
compute loss for weight  5e-06  0 result 0.43759
 training batch 29 mu var0-0.442964
compute loss for weight  -5e-06  0 result 0.43759
   --dy = -1.4803e-11 dy_ref = 1.04083e-17
 training batch 30 mu var0-0.442964
compute loss for weight  1e-05  0 result 0.43759
 training batch 31 mu var0-0.442964
compute loss for weight  -1e-05  0 result 0.43759
 training batch 32 mu var0-0.442964
compute loss for weight  5e-06  0 result 0.43759
 training batch 33 mu var0-0.442964
compute loss for weight  -5e-06  0 result 0.43759
   --dy = 1.85037e-12 dy_ref = -8.67362e-19
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.308     -0.6175 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.6194     -0.1053 

 training batch 34 mu var0-0.442964
compute loss for weight  -0.619418  -0.619428 result 0.437577
 training batch 35 mu var0-0.442964
compute loss for weight  -0.619438  -0.619428 result 0.437603
 training batch 36 mu var0-0.442964
compute loss for weight  -0.619423  -0.619428 result 0.437584
 training batch 37 mu var0-0.442964
compute loss for weight  -0.619433  -0.619428 result 0.437597
   --dy = -1.30792 dy_ref = -1.30792
 training batch 38 mu var0-0.442964
compute loss for weight  -0.105293  -0.105303 result 0.437584
 training batch 39 mu var0-0.442964
compute loss for weight  -0.105313  -0.105303 result 0.437596
 training batch 40 mu var0-0.442964
compute loss for weight  -0.105298  -0.105303 result 0.437587
 training batch 41 mu var0-0.442964
compute loss for weight  -0.105308  -0.105303 result 0.437593
   --dy = -0.617464 dy_ref = -0.617464
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m4.9591e-10[NON-XML-CHAR-0x1B][39m