Execution Time0.55s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: PR-4624-x86_64-fedora27-gcc7-opt (sft-fedora-27-2.cern.ch) on 2019-11-14 19:00:39
Repository revision: 164a856564302eda66fd0717d8545a574cef2806

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 var0-1.16079
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.742    -0.04458 
   1 |     -1.793     -0.2774 
   2 |     -1.651       1.262 
   3 |     -3.517       1.063 
   4 |     -2.215      0.2883 
   5 |     -4.614       0.182 
   6 |      3.283      -1.967 
   7 |     0.6979      0.6977 
   8 |     0.1186      -1.221 
   9 |    -0.1753     0.06525 

output BN 
output DL feature 0 mean -1.16079	output DL std 2.24529
output DL feature 1 mean 0.00471232	output DL std 0.987042
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.2729    -0.05264 
   1 |    -0.2967     -0.3013 
   2 |    -0.2302       1.342 
   3 |     -1.106        1.13 
   4 |    -0.4947      0.3028 
   5 |     -1.621      0.1893 
   6 |      2.086      -2.106 
   7 |     0.8726        0.74 
   8 |     0.6006      -1.309 
   9 |     0.4627     0.06465 

output BN feature 0 mean 3.88578e-17	output BN std 1.05408
output BN feature 1 mean -4.02456e-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 |   -0.03028     0.02153   -0.006567    -0.03297 
   1 |    -0.1177     0.07505    -0.02538     -0.1727 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     -1.672      0.5321      0.2748       1.829 
   1 |    -0.1932      0.1075      0.2099      -1.031 

 training batch 2 mu var0-1.16079
compute loss for weight  -1.67235  -1.67236 result 0.964535
 training batch 3 mu var0-1.16079
compute loss for weight  -1.67237  -1.67236 result 0.964535
 training batch 4 mu var0-1.16079
compute loss for weight  -1.67235  -1.67236 result 0.964535
 training batch 5 mu var0-1.16079
compute loss for weight  -1.67236  -1.67236 result 0.964535
   --dy = -0.0302845 dy_ref = -0.0302845
 training batch 6 mu var0-1.16079
compute loss for weight  0.532148  0.532138 result 0.964535
 training batch 7 mu var0-1.16079
compute loss for weight  0.532128  0.532138 result 0.964535
 training batch 8 mu var0-1.16079
compute loss for weight  0.532143  0.532138 result 0.964535
 training batch 9 mu var0-1.16079
compute loss for weight  0.532133  0.532138 result 0.964535
   --dy = 0.0215299 dy_ref = 0.0215299
 training batch 10 mu var0-1.16079
compute loss for weight  0.274832  0.274822 result 0.964535
 training batch 11 mu var0-1.16079
compute loss for weight  0.274812  0.274822 result 0.964535
 training batch 12 mu var0-1.16079
compute loss for weight  0.274827  0.274822 result 0.964535
 training batch 13 mu var0-1.16079
compute loss for weight  0.274817  0.274822 result 0.964535
   --dy = -0.00656657 dy_ref = -0.00656657
 training batch 14 mu var0-1.16079
compute loss for weight  1.82902  1.82901 result 0.964535
 training batch 15 mu var0-1.16079
compute loss for weight  1.829  1.82901 result 0.964535
 training batch 16 mu var0-1.16079
compute loss for weight  1.82901  1.82901 result 0.964535
 training batch 17 mu var0-1.16079
compute loss for weight  1.829  1.82901 result 0.964535
   --dy = -0.0329691 dy_ref = -0.0329691
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |   -0.08938       2.018 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-1.16079
compute loss for weight  1.00001  1 result 0.964534
 training batch 19 mu var0-1.16079
compute loss for weight  0.99999  1 result 0.964536
 training batch 20 mu var0-1.16079
compute loss for weight  1.00001  1 result 0.964535
 training batch 21 mu var0-1.16079
compute loss for weight  0.999995  1 result 0.964535
   --dy = -0.089379 dy_ref = -0.089379
 training batch 22 mu var0-1.16079
compute loss for weight  1.00001  1 result 0.964555
 training batch 23 mu var0-1.16079
compute loss for weight  0.99999  1 result 0.964515
 training batch 24 mu var0-1.16079
compute loss for weight  1.00001  1 result 0.964545
 training batch 25 mu var0-1.16079
compute loss for weight  0.999995  1 result 0.964525
   --dy = 2.01845 dy_ref = 2.01845
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  4.337e-19  -1.422e-16 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-1.16079
compute loss for weight  1e-05  0 result 0.964535
 training batch 27 mu var0-1.16079
compute loss for weight  -1e-05  0 result 0.964535
 training batch 28 mu var0-1.16079
compute loss for weight  5e-06  0 result 0.964535
 training batch 29 mu var0-1.16079
compute loss for weight  -5e-06  0 result 0.964535
   --dy = -3.88578e-11 dy_ref = 4.33681e-19
 training batch 30 mu var0-1.16079
compute loss for weight  1e-05  0 result 0.964535
 training batch 31 mu var0-1.16079
compute loss for weight  -1e-05  0 result 0.964535
 training batch 32 mu var0-1.16079
compute loss for weight  5e-06  0 result 0.964535
 training batch 33 mu var0-1.16079
compute loss for weight  -5e-06  0 result 0.964535
   --dy = 5.55112e-12 dy_ref = -1.42247e-16
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.168      -1.961 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |   -0.07654       -1.03 

 training batch 34 mu var0-1.16079
compute loss for weight  -0.0765317  -0.0765417 result 0.964547
 training batch 35 mu var0-1.16079
compute loss for weight  -0.0765517  -0.0765417 result 0.964523
 training batch 36 mu var0-1.16079
compute loss for weight  -0.0765367  -0.0765417 result 0.964541
 training batch 37 mu var0-1.16079
compute loss for weight  -0.0765467  -0.0765417 result 0.964529
   --dy = 1.16772 dy_ref = 1.16772
 training batch 38 mu var0-1.16079
compute loss for weight  -1.0295  -1.02951 result 0.964515
 training batch 39 mu var0-1.16079
compute loss for weight  -1.02952  -1.02951 result 0.964555
 training batch 40 mu var0-1.16079
compute loss for weight  -1.02951  -1.02951 result 0.964525
 training batch 41 mu var0-1.16079
compute loss for weight  -1.02952  -1.02951 result 0.964545
   --dy = -1.96059 dy_ref = -1.96059
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m1.30293e-09[NON-XML-CHAR-0x1B][39m