Execution Time0.55s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-mac1014-clang100-dbg (macphsft17.dyndns.cern.ch) on 2019-11-18 11:04:51

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 var00.345034
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.6972     -0.9071 
   1 |     0.1382     -0.6017 
   2 |     0.6243     -0.3756 
   3 |    -0.5035    -0.03885 
   4 |      0.493     -0.7455 
   5 |      1.274      -1.883 
   6 |    -0.7866      0.6264 
   7 |     0.6237    -0.07483 
   8 |      1.103        -1.2 
   9 |    -0.2127      0.1396 

output BN 
output DL feature 0 mean 0.345034	output DL std 0.674389
output DL feature 1 mean -0.505977	output DL std 0.724482
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.5505     -0.5836 
   1 |    -0.3233     -0.1393 
   2 |     0.4364      0.1897 
   3 |     -1.326      0.6796 
   4 |     0.2312     -0.3484 
   5 |      1.452      -2.003 
   6 |     -1.769       1.647 
   7 |     0.4355      0.6272 
   8 |      1.184      -1.009 
   9 |    -0.8717      0.9392 

output BN feature 0 mean -5.55112e-17	output BN std 1.05396
output BN feature 1 mean 7.77156e-17	output BN std 1.05398
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    0.05712    -0.08608     -0.0155   -0.007492 
   1 |      1.112     -0.3961      0.7789     -0.2092 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.5533      0.3231      0.3544     -0.2635 
   1 |    -0.8455    -0.08886     -0.1201      0.2419 

 training batch 2 mu var00.345037
compute loss for weight  0.553316  0.553306 result 0.645807
 training batch 3 mu var00.345034
compute loss for weight  0.553296  0.553306 result 0.645806
 training batch 4 mu var00.345034
compute loss for weight  0.553311  0.553306 result 0.645807
 training batch 5 mu var00.345034
compute loss for weight  0.553301  0.553306 result 0.645806
   --dy = 0.0571211 dy_ref = 0.0571211
 training batch 6 mu var00.345033
compute loss for weight  0.323081  0.323071 result 0.645806
 training batch 7 mu var00.345034
compute loss for weight  0.323061  0.323071 result 0.645807
 training batch 8 mu var00.345034
compute loss for weight  0.323076  0.323071 result 0.645806
 training batch 9 mu var00.345034
compute loss for weight  0.323066  0.323071 result 0.645807
   --dy = -0.0860832 dy_ref = -0.0860832
 training batch 10 mu var00.345034
compute loss for weight  0.354412  0.354402 result 0.645806
 training batch 11 mu var00.345034
compute loss for weight  0.354392  0.354402 result 0.645807
 training batch 12 mu var00.345034
compute loss for weight  0.354407  0.354402 result 0.645806
 training batch 13 mu var00.345034
compute loss for weight  0.354397  0.354402 result 0.645806
   --dy = -0.0155001 dy_ref = -0.0155001
 training batch 14 mu var00.345034
compute loss for weight  -0.263474  -0.263484 result 0.645806
 training batch 15 mu var00.345034
compute loss for weight  -0.263494  -0.263484 result 0.645806
 training batch 16 mu var00.345034
compute loss for weight  -0.263479  -0.263484 result 0.645806
 training batch 17 mu var00.345034
compute loss for weight  -0.263489  -0.263484 result 0.645806
   --dy = -0.00749222 dy_ref = -0.00749222
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.127      0.1644 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.345034
compute loss for weight  1.00001  1 result 0.645818
 training batch 19 mu var00.345034
compute loss for weight  0.99999  1 result 0.645795
 training batch 20 mu var00.345034
compute loss for weight  1.00001  1 result 0.645812
 training batch 21 mu var00.345034
compute loss for weight  0.999995  1 result 0.645801
   --dy = 1.1272 dy_ref = 1.1272
 training batch 22 mu var00.345034
compute loss for weight  1.00001  1 result 0.645808
 training batch 23 mu var00.345034
compute loss for weight  0.99999  1 result 0.645805
 training batch 24 mu var00.345034
compute loss for weight  1.00001  1 result 0.645807
 training batch 25 mu var00.345034
compute loss for weight  0.999995  1 result 0.645806
   --dy = 0.164412 dy_ref = 0.164412
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  2.776e-17  -3.469e-18 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.345034
compute loss for weight  1e-05  0 result 0.645806
 training batch 27 mu var00.345034
compute loss for weight  -1e-05  0 result 0.645806
 training batch 28 mu var00.345034
compute loss for weight  5e-06  0 result 0.645806
 training batch 29 mu var00.345034
compute loss for weight  -5e-06  0 result 0.645806
   --dy = 1.66533e-11 dy_ref = 2.77556e-17
 training batch 30 mu var00.345034
compute loss for weight  1e-05  0 result 0.645806
 training batch 31 mu var00.345034
compute loss for weight  -1e-05  0 result 0.645806
 training batch 32 mu var00.345034
compute loss for weight  5e-06  0 result 0.645806
 training batch 33 mu var00.345034
compute loss for weight  -5e-06  0 result 0.645806
   --dy = -3.70074e-12 dy_ref = -3.46945e-18
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.603      -1.455 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.7031      -0.113 

 training batch 34 mu var00.345034
compute loss for weight  0.703067  0.703057 result 0.645822
 training batch 35 mu var00.345034
compute loss for weight  0.703047  0.703057 result 0.64579
 training batch 36 mu var00.345034
compute loss for weight  0.703062  0.703057 result 0.645814
 training batch 37 mu var00.345034
compute loss for weight  0.703052  0.703057 result 0.645798
   --dy = 1.60329 dy_ref = 1.60329
 training batch 38 mu var00.345034
compute loss for weight  -0.113024  -0.113034 result 0.645792
 training batch 39 mu var00.345034
compute loss for weight  -0.113044  -0.113034 result 0.645821
 training batch 40 mu var00.345034
compute loss for weight  -0.113029  -0.113034 result 0.645799
 training batch 41 mu var00.345034
compute loss for weight  -0.113039  -0.113034 result 0.645814
   --dy = -1.45454 dy_ref = -1.45454
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m5.44081e-09[NON-XML-CHAR-0x1B][39m