Execution Time0.32s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: PR-4624-x86_64-mac1014-clang100-opt (macphsft17.dyndns.cern.ch) on 2019-11-14 19:01:56

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 var00.579328
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.9099     -0.5247 
   1 |     0.8245      -0.251 
   2 |     0.3385      -1.752 
   3 |      1.347     -0.2258 
   4 |      1.036     -0.5376 
   5 |      2.251      -1.178 
   6 |     -0.379       3.862 
   7 |     -1.159      -2.864 
   8 |     0.5398      0.3168 
   9 |    0.08395      0.2378 

output BN 
output DL feature 0 mean 0.579328	output DL std 0.943158
output DL feature 1 mean -0.291582	output DL std 1.74994
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.3695     -0.1404 
   1 |      0.274     0.02442 
   2 |    -0.2691     -0.8796 
   3 |     0.8574     0.03961 
   4 |     0.5105     -0.1482 
   5 |      1.869     -0.5336 
   6 |     -1.071       2.502 
   7 |     -1.942       -1.55 
   8 |   -0.04415      0.3665 
   9 |    -0.5536      0.3189 

output BN feature 0 mean 3.33067e-17	output BN std 1.05403
output BN feature 1 mean -1.66533e-17	output BN std 1.05407
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |  -0.005194    -0.05052    -0.02673     0.08797 
   1 |    0.08482       0.137      0.1338     -0.4191 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.7752     -0.5517     -0.1454     -0.3127 
   1 |     -0.666      -1.156     -0.3983       1.693 

 training batch 2 mu var00.579331
compute loss for weight  0.775177  0.775167 result 0.906576
 training batch 3 mu var00.579328
compute loss for weight  0.775157  0.775167 result 0.906576
 training batch 4 mu var00.579329
compute loss for weight  0.775172  0.775167 result 0.906576
 training batch 5 mu var00.579328
compute loss for weight  0.775162  0.775167 result 0.906576
   --dy = -0.00519427 dy_ref = -0.00519427
 training batch 6 mu var00.579328
compute loss for weight  -0.55165  -0.55166 result 0.906576
 training batch 7 mu var00.579328
compute loss for weight  -0.55167  -0.55166 result 0.906577
 training batch 8 mu var00.579328
compute loss for weight  -0.551655  -0.55166 result 0.906576
 training batch 9 mu var00.579328
compute loss for weight  -0.551665  -0.55166 result 0.906577
   --dy = -0.0505238 dy_ref = -0.0505238
 training batch 10 mu var00.579329
compute loss for weight  -0.145352  -0.145362 result 0.906576
 training batch 11 mu var00.579328
compute loss for weight  -0.145372  -0.145362 result 0.906577
 training batch 12 mu var00.579328
compute loss for weight  -0.145357  -0.145362 result 0.906576
 training batch 13 mu var00.579328
compute loss for weight  -0.145367  -0.145362 result 0.906577
   --dy = -0.0267328 dy_ref = -0.0267328
 training batch 14 mu var00.579328
compute loss for weight  -0.312666  -0.312676 result 0.906577
 training batch 15 mu var00.579328
compute loss for weight  -0.312686  -0.312676 result 0.906576
 training batch 16 mu var00.579328
compute loss for weight  -0.312671  -0.312676 result 0.906577
 training batch 17 mu var00.579328
compute loss for weight  -0.312681  -0.312676 result 0.906576
   --dy = 0.0879665 dy_ref = 0.0879665
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.814   -0.000622 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.579328
compute loss for weight  1.00001  1 result 0.906595
 training batch 19 mu var00.579328
compute loss for weight  0.99999  1 result 0.906558
 training batch 20 mu var00.579328
compute loss for weight  1.00001  1 result 0.906585
 training batch 21 mu var00.579328
compute loss for weight  0.999995  1 result 0.906567
   --dy = 1.81377 dy_ref = 1.81377
 training batch 22 mu var00.579328
compute loss for weight  1.00001  1 result 0.906576
 training batch 23 mu var00.579328
compute loss for weight  0.99999  1 result 0.906576
 training batch 24 mu var00.579328
compute loss for weight  1.00001  1 result 0.906576
 training batch 25 mu var00.579328
compute loss for weight  0.999995  1 result 0.906576
   --dy = -0.000622028 dy_ref = -0.000622028
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  8.327e-17   8.674e-18 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.579328
compute loss for weight  1e-05  0 result 0.906576
 training batch 27 mu var00.579328
compute loss for weight  -1e-05  0 result 0.906576
 training batch 28 mu var00.579328
compute loss for weight  5e-06  0 result 0.906576
 training batch 29 mu var00.579328
compute loss for weight  -5e-06  0 result 0.906576
   --dy = -4.44089e-11 dy_ref = 8.32667e-17
 training batch 30 mu var00.579328
compute loss for weight  1e-05  0 result 0.906576
 training batch 31 mu var00.579328
compute loss for weight  -1e-05  0 result 0.906576
 training batch 32 mu var00.579328
compute loss for weight  5e-06  0 result 0.906576
 training batch 33 mu var00.579328
compute loss for weight  -5e-06  0 result 0.906576
   --dy = 3.33067e-11 dy_ref = 8.67362e-18
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |       -1.9    0.009898 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.9546    -0.06284 

 training batch 34 mu var00.579328
compute loss for weight  -0.95459  -0.9546 result 0.906557
 training batch 35 mu var00.579328
compute loss for weight  -0.95461  -0.9546 result 0.906595
 training batch 36 mu var00.579328
compute loss for weight  -0.954595  -0.9546 result 0.906567
 training batch 37 mu var00.579328
compute loss for weight  -0.954605  -0.9546 result 0.906586
   --dy = -1.90004 dy_ref = -1.90004
 training batch 38 mu var00.579328
compute loss for weight  -0.0628335  -0.0628435 result 0.906576
 training batch 39 mu var00.579328
compute loss for weight  -0.0628535  -0.0628435 result 0.906576
 training batch 40 mu var00.579328
compute loss for weight  -0.0628385  -0.0628435 result 0.906576
 training batch 41 mu var00.579328
compute loss for weight  -0.0628485  -0.0628435 result 0.906576
   --dy = 0.00989805 dy_ref = 0.00989805
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m1.80584e-08[NON-XML-CHAR-0x1B][39m