Execution Time0.30s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: PR-4624-x86_64-mac1014-clang100-opt (macphsft17.dyndns.cern.ch) on 2019-11-14 18:11:27

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 var0-0.556673
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.004     -0.6637 
   1 |    -0.5827      -3.037 
   2 |    0.07527       3.325 
   3 |   -0.03858      0.2531 
   4 |    -0.8198    -0.05987 
   5 |     -2.054      -1.581 
   6 |    -0.7545      -1.774 
   7 |      1.213      0.3833 
   8 |     -1.679      -2.034 
   9 |    0.07858    -0.00773 

output BN 
output DL feature 0 mean -0.556673	output DL std 0.943712
output DL feature 1 mean -0.519637	output DL std 1.76435
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.5001    -0.08609 
   1 |   -0.02909      -1.504 
   2 |     0.7058       2.297 
   3 |     0.5787      0.4617 
   4 |    -0.2939      0.2747 
   5 |     -1.672     -0.6341 
   6 |    -0.2209     -0.7496 
   7 |      1.977      0.5394 
   8 |     -1.254      -0.905 
   9 |     0.7095      0.3058 

output BN feature 0 mean -5.55112e-17	output BN std 1.05403
output BN feature 1 mean -3.88578e-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 |   0.008732      0.1187     -0.2119      0.1654 
   1 |     0.1482     -0.1307      0.1208     0.05583 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    -0.7767      0.4385    -0.06212     -0.3532 
   1 |     -2.061     -0.2806       1.588      -1.032 

 training batch 2 mu var0-0.55667
compute loss for weight  -0.776685  -0.776695 result 0.195856
 training batch 3 mu var0-0.556673
compute loss for weight  -0.776705  -0.776695 result 0.195856
 training batch 4 mu var0-0.556672
compute loss for weight  -0.77669  -0.776695 result 0.195856
 training batch 5 mu var0-0.556673
compute loss for weight  -0.7767  -0.776695 result 0.195856
   --dy = 0.00873212 dy_ref = 0.00873212
 training batch 6 mu var0-0.556673
compute loss for weight  0.438465  0.438455 result 0.195857
 training batch 7 mu var0-0.556673
compute loss for weight  0.438445  0.438455 result 0.195855
 training batch 8 mu var0-0.556673
compute loss for weight  0.43846  0.438455 result 0.195857
 training batch 9 mu var0-0.556673
compute loss for weight  0.43845  0.438455 result 0.195855
   --dy = 0.118747 dy_ref = 0.118747
 training batch 10 mu var0-0.556672
compute loss for weight  -0.0621092  -0.0621192 result 0.195854
 training batch 11 mu var0-0.556673
compute loss for weight  -0.0621292  -0.0621192 result 0.195858
 training batch 12 mu var0-0.556673
compute loss for weight  -0.0621142  -0.0621192 result 0.195855
 training batch 13 mu var0-0.556673
compute loss for weight  -0.0621242  -0.0621192 result 0.195857
   --dy = -0.211915 dy_ref = -0.211915
 training batch 14 mu var0-0.556673
compute loss for weight  -0.3532  -0.35321 result 0.195858
 training batch 15 mu var0-0.556673
compute loss for weight  -0.35322  -0.35321 result 0.195854
 training batch 16 mu var0-0.556673
compute loss for weight  -0.353205  -0.35321 result 0.195857
 training batch 17 mu var0-0.556673
compute loss for weight  -0.353215  -0.35321 result 0.195855
   --dy = 0.165385 dy_ref = 0.165385
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2516      0.1401 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.556673
compute loss for weight  1.00001  1 result 0.195859
 training batch 19 mu var0-0.556673
compute loss for weight  0.99999  1 result 0.195853
 training batch 20 mu var0-0.556673
compute loss for weight  1.00001  1 result 0.195857
 training batch 21 mu var0-0.556673
compute loss for weight  0.999995  1 result 0.195855
   --dy = 0.25164 dy_ref = 0.25164
 training batch 22 mu var0-0.556673
compute loss for weight  1.00001  1 result 0.195857
 training batch 23 mu var0-0.556673
compute loss for weight  0.99999  1 result 0.195855
 training batch 24 mu var0-0.556673
compute loss for weight  1.00001  1 result 0.195857
 training batch 25 mu var0-0.556673
compute loss for weight  0.999995  1 result 0.195855
   --dy = 0.140072 dy_ref = 0.140072
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -2.776e-17   1.041e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.556673
compute loss for weight  1e-05  0 result 0.195856
 training batch 27 mu var0-0.556673
compute loss for weight  -1e-05  0 result 0.195856
 training batch 28 mu var0-0.556673
compute loss for weight  5e-06  0 result 0.195856
 training batch 29 mu var0-0.556673
compute loss for weight  -5e-06  0 result 0.195856
   --dy = -4.62593e-13 dy_ref = -2.77556e-17
 training batch 30 mu var0-0.556673
compute loss for weight  1e-05  0 result 0.195856
 training batch 31 mu var0-0.556673
compute loss for weight  -1e-05  0 result 0.195856
 training batch 32 mu var0-0.556673
compute loss for weight  5e-06  0 result 0.195856
 training batch 33 mu var0-0.556673
compute loss for weight  -5e-06  0 result 0.195856
   --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 |    -0.5082      0.3219 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.4952      0.4352 

 training batch 34 mu var0-0.556673
compute loss for weight  -0.495173  -0.495183 result 0.195851
 training batch 35 mu var0-0.556673
compute loss for weight  -0.495193  -0.495183 result 0.195861
 training batch 36 mu var0-0.556673
compute loss for weight  -0.495178  -0.495183 result 0.195853
 training batch 37 mu var0-0.556673
compute loss for weight  -0.495188  -0.495183 result 0.195859
   --dy = -0.508175 dy_ref = -0.508175
 training batch 38 mu var0-0.556673
compute loss for weight  0.43521  0.4352 result 0.195859
 training batch 39 mu var0-0.556673
compute loss for weight  0.43519  0.4352 result 0.195853
 training batch 40 mu var0-0.556673
compute loss for weight  0.435205  0.4352 result 0.195858
 training batch 41 mu var0-0.556673
compute loss for weight  0.435195  0.4352 result 0.195854
   --dy = 0.321857 dy_ref = 0.321857
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m6.23618e-10[NON-XML-CHAR-0x1B][39m