Execution Time0.20s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: PR-420-i686-ubuntu18-gcc7-opt (sft-ubuntu-1804-i386-2) on 2019-11-14 10:12:12
Repository revision: 56f0d5eaeec4ca0a503d7668700eb66974f83206

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 var00.00619006
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    0.04751      -1.025 
   1 |     0.4581     -0.3668 
   2 |     -1.066      -1.107 
   3 |     -1.075       -1.47 
   4 |    -0.3045      -1.245 
   5 |    -0.1452      -2.463 
   6 |      1.009      0.8353 
   7 |     0.2309       1.182 
   8 |      1.023     -0.4399 
   9 |    -0.1167    -0.05759 

output BN 
output DL feature 0 mean 0.00619006	output DL std 0.726584
output DL feature 1 mean -0.615639	output DL std 1.08986
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    0.05994     -0.3961 
   1 |     0.6556      0.2406 
   2 |     -1.555     -0.4754 
   3 |     -1.569     -0.8258 
   4 |    -0.4507     -0.6084 
   5 |    -0.2197      -1.787 
   6 |      1.455       1.403 
   7 |      0.326       1.739 
   8 |      1.476        0.17 
   9 |    -0.1782      0.5397 

output BN feature 0 mean 1.66533e-17	output BN std 1.05398
output BN feature 1 mean 6.66134e-17	output BN std 1.05404
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    -0.7062      0.5974     -0.3855   0.0009113 
   1 |     0.7962     -0.6462      0.4317     0.04313 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.3781      0.2958     -0.2336       0.632 
   1 |     -0.571      0.6427     -0.2026      0.5652 

 training batch 2 mu var00.00619288
compute loss for weight  0.378155  0.378145 result 1.31564
 training batch 3 mu var00.00619006
compute loss for weight  0.378135  0.378145 result 1.31565
 training batch 4 mu var00.00619077
compute loss for weight  0.37815  0.378145 result 1.31564
 training batch 5 mu var00.00619006
compute loss for weight  0.37814  0.378145 result 1.31565
   --dy = -0.7062 dy_ref = -0.7062
 training batch 6 mu var00.00618957
compute loss for weight  0.295821  0.295811 result 1.31565
 training batch 7 mu var00.00619006
compute loss for weight  0.295801  0.295811 result 1.31564
 training batch 8 mu var00.00618988
compute loss for weight  0.295816  0.295811 result 1.31565
 training batch 9 mu var00.00619006
compute loss for weight  0.295806  0.295811 result 1.31564
   --dy = 0.597361 dy_ref = 0.597361
 training batch 10 mu var00.00619036
compute loss for weight  -0.233592  -0.233602 result 1.31564
 training batch 11 mu var00.00619006
compute loss for weight  -0.233612  -0.233602 result 1.31565
 training batch 12 mu var00.00619019
compute loss for weight  -0.233597  -0.233602 result 1.31564
 training batch 13 mu var00.00619006
compute loss for weight  -0.233607  -0.233602 result 1.31564
   --dy = -0.385521 dy_ref = -0.385521
 training batch 14 mu var00.00619
compute loss for weight  0.632025  0.632015 result 1.31564
 training batch 15 mu var00.00619006
compute loss for weight  0.632005  0.632015 result 1.31564
 training batch 16 mu var00.00619004
compute loss for weight  0.63202  0.632015 result 1.31564
 training batch 17 mu var00.00619006
compute loss for weight  0.63201  0.632015 result 1.31564
   --dy = 0.000911289 dy_ref = 0.000911289
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |        1.4       1.232 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.00619006
compute loss for weight  1.00001  1 result 1.31566
 training batch 19 mu var00.00619006
compute loss for weight  0.99999  1 result 1.31563
 training batch 20 mu var00.00619006
compute loss for weight  1.00001  1 result 1.31565
 training batch 21 mu var00.00619006
compute loss for weight  0.999995  1 result 1.31564
   --dy = 1.39976 dy_ref = 1.39976
 training batch 22 mu var00.00619006
compute loss for weight  1.00001  1 result 1.31566
 training batch 23 mu var00.00619006
compute loss for weight  0.99999  1 result 1.31563
 training batch 24 mu var00.00619006
compute loss for weight  1.00001  1 result 1.31565
 training batch 25 mu var00.00619006
compute loss for weight  0.999995  1 result 1.31564
   --dy = 1.23152 dy_ref = 1.23152
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  3.469e-18  -1.041e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.00619006
compute loss for weight  1e-05  0 result 1.31564
 training batch 27 mu var00.00619006
compute loss for weight  -1e-05  0 result 1.31564
 training batch 28 mu var00.00619006
compute loss for weight  5e-06  0 result 1.31564
 training batch 29 mu var00.00619006
compute loss for weight  -5e-06  0 result 1.31564
   --dy = 0 dy_ref = 3.46945e-18
 training batch 30 mu var00.00619006
compute loss for weight  1e-05  0 result 1.31564
 training batch 31 mu var00.00619006
compute loss for weight  -1e-05  0 result 1.31564
 training batch 32 mu var00.00619006
compute loss for weight  5e-06  0 result 1.31564
 training batch 33 mu var00.00619006
compute loss for weight  -5e-06  0 result 1.31564
   --dy = -3.70074e-12 dy_ref = -1.04083e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      2.054       1.997 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.6816      0.6168 

 training batch 34 mu var00.00619006
compute loss for weight  0.68162  0.68161 result 1.31566
 training batch 35 mu var00.00619006
compute loss for weight  0.6816  0.68161 result 1.31562
 training batch 36 mu var00.00619006
compute loss for weight  0.681615  0.68161 result 1.31565
 training batch 37 mu var00.00619006
compute loss for weight  0.681605  0.68161 result 1.31563
   --dy = 2.05361 dy_ref = 2.05361
 training batch 38 mu var00.00619006
compute loss for weight  0.616783  0.616773 result 1.31566
 training batch 39 mu var00.00619006
compute loss for weight  0.616763  0.616773 result 1.31562
 training batch 40 mu var00.00619006
compute loss for weight  0.616778  0.616773 result 1.31565
 training batch 41 mu var00.00619006
compute loss for weight  0.616768  0.616773 result 1.31563
   --dy = 1.99672 dy_ref = 1.99672
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m1.07688e-07[NON-XML-CHAR-0x1B][39m