Execution Time0.51s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-mac1015-clang110 (macphsft19.dyndns.cern.ch) on 2019-11-15 01:13:00

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 var00.158128
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      0.404      -0.587 
   1 |    -0.1896     -0.1013 
   2 |     0.5247      0.4946 
   3 |    -0.8606       1.959 
   4 |     0.1853     -0.0156 
   5 |     0.5587     -0.6477 
   6 |    -0.4866      -1.097 
   7 |     0.7463     -0.3078 
   8 |      0.918      -2.262 
   9 |    -0.2188      0.3285 

output BN 
output DL feature 0 mean 0.158128	output DL std 0.576958
output DL feature 1 mean -0.223633	output DL std 1.09949
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.4491     -0.3483 
   1 |    -0.6353      0.1173 
   2 |     0.6696      0.6886 
   3 |     -1.861       2.092 
   4 |    0.04962      0.1994 
   5 |     0.7317     -0.4066 
   6 |     -1.178     -0.8375 
   7 |      1.074    -0.08068 
   8 |      1.388      -1.954 
   9 |    -0.6886      0.5293 

output BN feature 0 mean 8.88178e-17	output BN std 1.05392
output BN feature 1 mean -1.11022e-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 |     0.1429     0.01696     0.04621      0.5759 
   1 |     0.8023      0.4984      0.8892       1.762 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.2414      0.3805      0.4159      -0.102 
   1 |     -0.544     -0.3819     -0.2976     -0.7676 

 training batch 2 mu var00.15813
compute loss for weight  0.241446  0.241436 result 2.30226
 training batch 3 mu var00.158128
compute loss for weight  0.241426  0.241436 result 2.30226
 training batch 4 mu var00.158128
compute loss for weight  0.241441  0.241436 result 2.30226
 training batch 5 mu var00.158128
compute loss for weight  0.241431  0.241436 result 2.30226
   --dy = 0.142856 dy_ref = 0.142856
 training batch 6 mu var00.158127
compute loss for weight  0.380525  0.380515 result 2.30226
 training batch 7 mu var00.158128
compute loss for weight  0.380505  0.380515 result 2.30226
 training batch 8 mu var00.158127
compute loss for weight  0.38052  0.380515 result 2.30226
 training batch 9 mu var00.158128
compute loss for weight  0.38051  0.380515 result 2.30226
   --dy = 0.0169569 dy_ref = 0.0169569
 training batch 10 mu var00.158128
compute loss for weight  0.415926  0.415916 result 2.30226
 training batch 11 mu var00.158128
compute loss for weight  0.415906  0.415916 result 2.30226
 training batch 12 mu var00.158128
compute loss for weight  0.415921  0.415916 result 2.30226
 training batch 13 mu var00.158128
compute loss for weight  0.415911  0.415916 result 2.30226
   --dy = 0.0462096 dy_ref = 0.0462096
 training batch 14 mu var00.158128
compute loss for weight  -0.101975  -0.101985 result 2.30227
 training batch 15 mu var00.158128
compute loss for weight  -0.101995  -0.101985 result 2.30225
 training batch 16 mu var00.158128
compute loss for weight  -0.10198  -0.101985 result 2.30226
 training batch 17 mu var00.158128
compute loss for weight  -0.10199  -0.101985 result 2.30226
   --dy = 0.575922 dy_ref = 0.575922
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      4.276      0.3286 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.158128
compute loss for weight  1.00001  1 result 2.3023
 training batch 19 mu var00.158128
compute loss for weight  0.99999  1 result 2.30222
 training batch 20 mu var00.158128
compute loss for weight  1.00001  1 result 2.30228
 training batch 21 mu var00.158128
compute loss for weight  0.999995  1 result 2.30224
   --dy = 4.27592 dy_ref = 4.27592
 training batch 22 mu var00.158128
compute loss for weight  1.00001  1 result 2.30226
 training batch 23 mu var00.158128
compute loss for weight  0.99999  1 result 2.30226
 training batch 24 mu var00.158128
compute loss for weight  1.00001  1 result 2.30226
 training batch 25 mu var00.158128
compute loss for weight  0.999995  1 result 2.30226
   --dy = 0.328604 dy_ref = 0.328604
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  3.886e-16  -5.551e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.158128
compute loss for weight  1e-05  0 result 2.30226
 training batch 27 mu var00.158128
compute loss for weight  -1e-05  0 result 2.30226
 training batch 28 mu var00.158128
compute loss for weight  5e-06  0 result 2.30226
 training batch 29 mu var00.158128
compute loss for weight  -5e-06  0 result 2.30226
   --dy = 5.92119e-11 dy_ref = 3.88578e-16
 training batch 30 mu var00.158128
compute loss for weight  1e-05  0 result 2.30226
 training batch 31 mu var00.158128
compute loss for weight  -1e-05  0 result 2.30226
 training batch 32 mu var00.158128
compute loss for weight  5e-06  0 result 2.30226
 training batch 33 mu var00.158128
compute loss for weight  -5e-06  0 result 2.30226
   --dy = 0 dy_ref = -5.55112e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      3.024      -2.053 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.414       -0.16 

 training batch 34 mu var00.158128
compute loss for weight  1.4142  1.41419 result 2.30229
 training batch 35 mu var00.158128
compute loss for weight  1.41418  1.41419 result 2.30223
 training batch 36 mu var00.158128
compute loss for weight  1.41419  1.41419 result 2.30227
 training batch 37 mu var00.158128
compute loss for weight  1.41418  1.41419 result 2.30224
   --dy = 3.02358 dy_ref = 3.02358
 training batch 38 mu var00.158128
compute loss for weight  -0.160019  -0.160029 result 2.30224
 training batch 39 mu var00.158128
compute loss for weight  -0.160039  -0.160029 result 2.30228
 training batch 40 mu var00.158128
compute loss for weight  -0.160024  -0.160029 result 2.30225
 training batch 41 mu var00.158128
compute loss for weight  -0.160034  -0.160029 result 2.30227
   --dy = -2.0534 dy_ref = -2.0534
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m9.21291e-09[NON-XML-CHAR-0x1B][39m