Execution Time0.07s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-fedora28-gcc8 (sft-fedora-28-1.cern.ch) on 2019-11-15 01:10:47

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 var00.913408
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.633      0.1579 
   1 |   -0.03148     -0.2733 
   2 |      1.921        2.21 
   3 |      1.651       2.244 
   4 |      1.895      0.7792 
   5 |      3.661       0.908 
   6 |    -0.5256      -3.302 
   7 |      -2.29      0.6825 
   8 |      1.202      -1.959 
   9 |     0.0172      0.1543 

output BN 
output DL feature 0 mean 0.913408	output DL std 1.65665
output DL feature 1 mean 0.160199	output DL std 1.71344
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      0.458   -0.001393 
   1 |    -0.6012     -0.2667 
   2 |     0.6414       1.261 
   3 |     0.4693       1.282 
   4 |     0.6243      0.3808 
   5 |      1.748        0.46 
   6 |    -0.9156       -2.13 
   7 |     -2.038      0.3213 
   8 |     0.1837      -1.303 
   9 |    -0.5702   -0.003644 

output BN feature 0 mean 1.44329e-16	output BN std 1.05407
output BN feature 1 mean -1.60028e-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 |   -0.07702     0.02963    -0.05144       -0.23 
   1 |     0.1603     -0.1609       0.125      0.1446 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.5753      -1.183       0.677     -0.4965 
   1 |     -0.161    -0.05633      0.3077      -1.787 

 training batch 2 mu var00.913411
compute loss for weight  0.575287  0.575277 result 1.51635
 training batch 3 mu var00.913408
compute loss for weight  0.575267  0.575277 result 1.51636
 training batch 4 mu var00.913409
compute loss for weight  0.575282  0.575277 result 1.51635
 training batch 5 mu var00.913408
compute loss for weight  0.575272  0.575277 result 1.51636
   --dy = -0.0770185 dy_ref = -0.0770185
 training batch 6 mu var00.913408
compute loss for weight  -1.18324  -1.18325 result 1.51636
 training batch 7 mu var00.913408
compute loss for weight  -1.18326  -1.18325 result 1.51635
 training batch 8 mu var00.913408
compute loss for weight  -1.18325  -1.18325 result 1.51636
 training batch 9 mu var00.913408
compute loss for weight  -1.18326  -1.18325 result 1.51635
   --dy = 0.0296282 dy_ref = 0.0296282
 training batch 10 mu var00.913408
compute loss for weight  0.676981  0.676971 result 1.51635
 training batch 11 mu var00.913408
compute loss for weight  0.676961  0.676971 result 1.51636
 training batch 12 mu var00.913408
compute loss for weight  0.676976  0.676971 result 1.51635
 training batch 13 mu var00.913408
compute loss for weight  0.676966  0.676971 result 1.51636
   --dy = -0.051438 dy_ref = -0.051438
 training batch 14 mu var00.913408
compute loss for weight  -0.496509  -0.496519 result 1.51635
 training batch 15 mu var00.913408
compute loss for weight  -0.496529  -0.496519 result 1.51636
 training batch 16 mu var00.913408
compute loss for weight  -0.496514  -0.496519 result 1.51635
 training batch 17 mu var00.913408
compute loss for weight  -0.496524  -0.496519 result 1.51636
   --dy = -0.229994 dy_ref = -0.229994
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2449       2.788 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.913408
compute loss for weight  1.00001  1 result 1.51636
 training batch 19 mu var00.913408
compute loss for weight  0.99999  1 result 1.51635
 training batch 20 mu var00.913408
compute loss for weight  1.00001  1 result 1.51636
 training batch 21 mu var00.913408
compute loss for weight  0.999995  1 result 1.51635
   --dy = 0.244914 dy_ref = 0.244914
 training batch 22 mu var00.913408
compute loss for weight  1.00001  1 result 1.51638
 training batch 23 mu var00.913408
compute loss for weight  0.99999  1 result 1.51633
 training batch 24 mu var00.913408
compute loss for weight  1.00001  1 result 1.51637
 training batch 25 mu var00.913408
compute loss for weight  0.999995  1 result 1.51634
   --dy = 2.7878 dy_ref = 2.7878
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -4.337e-18  -8.674e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.913408
compute loss for weight  1e-05  0 result 1.51636
 training batch 27 mu var00.913408
compute loss for weight  -1e-05  0 result 1.51636
 training batch 28 mu var00.913408
compute loss for weight  5e-06  0 result 1.51636
 training batch 29 mu var00.913408
compute loss for weight  -5e-06  0 result 1.51636
   --dy = 2.96059e-11 dy_ref = -4.33681e-18
 training batch 30 mu var00.913408
compute loss for weight  1e-05  0 result 1.51636
 training batch 31 mu var00.913408
compute loss for weight  -1e-05  0 result 1.51636
 training batch 32 mu var00.913408
compute loss for weight  5e-06  0 result 1.51636
 training batch 33 mu var00.913408
compute loss for weight  -5e-06  0 result 1.51636
   --dy = 3.33067e-11 dy_ref = -8.67362e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.236      -2.435 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.1981      -1.145 

 training batch 34 mu var00.913408
compute loss for weight  -0.198069  -0.198079 result 1.51634
 training batch 35 mu var00.913408
compute loss for weight  -0.198089  -0.198079 result 1.51637
 training batch 36 mu var00.913408
compute loss for weight  -0.198074  -0.198079 result 1.51635
 training batch 37 mu var00.913408
compute loss for weight  -0.198084  -0.198079 result 1.51636
   --dy = -1.23645 dy_ref = -1.23645
 training batch 38 mu var00.913408
compute loss for weight  -1.14486  -1.14487 result 1.51633
 training batch 39 mu var00.913408
compute loss for weight  -1.14488  -1.14487 result 1.51638
 training batch 40 mu var00.913408
compute loss for weight  -1.14486  -1.14487 result 1.51634
 training batch 41 mu var00.913408
compute loss for weight  -1.14487  -1.14487 result 1.51637
   --dy = -2.43504 dy_ref = -2.43504
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m4.67516e-10[NON-XML-CHAR-0x1B][39m