Execution Time0.31s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-ubuntu18-gcc7 (sft-ubuntu-1804-3) on 2019-11-15 02:17:26
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 var0-0.0268471
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1478      0.4912 
   1 |    -0.8525     -0.4497 
   2 |     0.5055        1.71 
   3 |     -1.163      0.2547 
   4 |   -0.03409      0.5977 
   5 |    -0.1189       1.011 
   6 |     0.5774      -2.008 
   7 |   -0.03504       0.914 
   8 |      0.889      -0.034 
   9 |    -0.1846     -0.1385 

output BN 
output DL feature 0 mean -0.0268471	output DL std 0.625724
output DL feature 1 mean 0.23478	output DL std 1.0082
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2942      0.2681 
   1 |     -1.391     -0.7156 
   2 |     0.8967       1.542 
   3 |     -1.914     0.02084 
   4 |    -0.0122      0.3794 
   5 |    -0.1551      0.8115 
   6 |      1.018      -2.345 
   7 |    -0.0138      0.7101 
   8 |      1.543      -0.281 
   9 |    -0.2657     -0.3903 

output BN feature 0 mean 1.11022e-17	output BN std 1.05394
output BN feature 1 mean -3.88578e-17	output BN std 1.05403
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |      1.029      0.5026       3.195      -4.681 
   1 |     0.2536      0.8431        2.16       3.802 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    -0.3208     0.01716      0.6222      0.3558 
   1 |    0.08138      0.2656      0.6356     -0.9481 

 training batch 2 mu var0-0.0268443
compute loss for weight  -0.320821  -0.320831 result 4.06307
 training batch 3 mu var0-0.0268471
compute loss for weight  -0.320841  -0.320831 result 4.06305
 training batch 4 mu var0-0.0268464
compute loss for weight  -0.320826  -0.320831 result 4.06307
 training batch 5 mu var0-0.0268471
compute loss for weight  -0.320836  -0.320831 result 4.06306
   --dy = 1.02938 dy_ref = 1.02938
 training batch 6 mu var0-0.0268476
compute loss for weight  0.0171715  0.0171615 result 4.06307
 training batch 7 mu var0-0.0268471
compute loss for weight  0.0171515  0.0171615 result 4.06306
 training batch 8 mu var0-0.0268473
compute loss for weight  0.0171665  0.0171615 result 4.06306
 training batch 9 mu var0-0.0268471
compute loss for weight  0.0171565  0.0171615 result 4.06306
   --dy = 0.502635 dy_ref = 0.502635
 training batch 10 mu var0-0.0268468
compute loss for weight  0.62222  0.62221 result 4.06309
 training batch 11 mu var0-0.0268471
compute loss for weight  0.6222  0.62221 result 4.06303
 training batch 12 mu var0-0.026847
compute loss for weight  0.622215  0.62221 result 4.06308
 training batch 13 mu var0-0.0268471
compute loss for weight  0.622205  0.62221 result 4.06305
   --dy = 3.19533 dy_ref = 3.19533
 training batch 14 mu var0-0.0268472
compute loss for weight  0.35581  0.3558 result 4.06301
 training batch 15 mu var0-0.0268471
compute loss for weight  0.35579  0.3558 result 4.06311
 training batch 16 mu var0-0.0268472
compute loss for weight  0.355805  0.3558 result 4.06304
 training batch 17 mu var0-0.0268471
compute loss for weight  0.355795  0.3558 result 4.06309
   --dy = -4.68136 dy_ref = -4.68136
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      3.204       4.922 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.0268471
compute loss for weight  1.00001  1 result 4.06309
 training batch 19 mu var0-0.0268471
compute loss for weight  0.99999  1 result 4.06303
 training batch 20 mu var0-0.0268471
compute loss for weight  1.00001  1 result 4.06308
 training batch 21 mu var0-0.0268471
compute loss for weight  0.999995  1 result 4.06305
   --dy = 3.20366 dy_ref = 3.20366
 training batch 22 mu var0-0.0268471
compute loss for weight  1.00001  1 result 4.06311
 training batch 23 mu var0-0.0268471
compute loss for weight  0.99999  1 result 4.06301
 training batch 24 mu var0-0.0268471
compute loss for weight  1.00001  1 result 4.06309
 training batch 25 mu var0-0.0268471
compute loss for weight  0.999995  1 result 4.06304
   --dy = 4.92247 dy_ref = 4.92247
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -1.665e-16    1.11e-16 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.0268471
compute loss for weight  1e-05  0 result 4.06306
 training batch 27 mu var0-0.0268471
compute loss for weight  -1e-05  0 result 4.06306
 training batch 28 mu var0-0.0268471
compute loss for weight  5e-06  0 result 4.06306
 training batch 29 mu var0-0.0268471
compute loss for weight  -5e-06  0 result 4.06306
   --dy = 0 dy_ref = -1.66533e-16
 training batch 30 mu var0-0.0268471
compute loss for weight  1e-05  0 result 4.06306
 training batch 31 mu var0-0.0268471
compute loss for weight  -1e-05  0 result 4.06306
 training batch 32 mu var0-0.0268471
compute loss for weight  5e-06  0 result 4.06306
 training batch 33 mu var0-0.0268471
compute loss for weight  -5e-06  0 result 4.06306
   --dy = 1.18424e-10 dy_ref = 1.11022e-16
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      2.462       3.081 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.301       1.598 

 training batch 34 mu var0-0.0268471
compute loss for weight  1.30144  1.30143 result 4.06309
 training batch 35 mu var0-0.0268471
compute loss for weight  1.30142  1.30143 result 4.06304
 training batch 36 mu var0-0.0268471
compute loss for weight  1.30143  1.30143 result 4.06307
 training batch 37 mu var0-0.0268471
compute loss for weight  1.30142  1.30143 result 4.06305
   --dy = 2.46165 dy_ref = 2.46165
 training batch 38 mu var0-0.0268471
compute loss for weight  1.59779  1.59778 result 4.06309
 training batch 39 mu var0-0.0268471
compute loss for weight  1.59777  1.59778 result 4.06303
 training batch 40 mu var0-0.0268471
compute loss for weight  1.59779  1.59778 result 4.06308
 training batch 41 mu var0-0.0268471
compute loss for weight  1.59778  1.59778 result 4.06305
   --dy = 3.08081 dy_ref = 3.08081
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m3.1822e-10[NON-XML-CHAR-0x1B][39m