Execution Time0.31s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-i686-ubuntu18-gcc7 (sft-ubuntu-1804-i386-2) on 2019-11-14 01:49:35
Repository revision: 32b17abcda23e44b64218a42d0ca69cb30cda7e0

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 var00.245253
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1623    -0.09391 
   1 |     0.9165     0.07065 
   2 |   -0.01423     -0.4971 
   3 |      2.395      0.5738 
   4 |     0.6111     0.03286 
   5 |      1.061    -0.04406 
   6 |    -0.6689       1.195 
   7 |     -1.162      -1.422 
   8 |     -1.183     -0.1563 
   9 |     0.3351      0.1687 

output BN 
output DL feature 0 mean 0.245253	output DL std 1.09601
output DL feature 1 mean -0.0172857	output DL std 0.674362
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.0798     -0.1198 
   1 |     0.6456      0.1374 
   2 |    -0.2495     -0.7498 
   3 |      2.067      0.9238 
   4 |     0.3518     0.07837 
   5 |     0.7846    -0.04185 
   6 |    -0.8792       1.894 
   7 |     -1.354      -2.196 
   8 |     -1.374     -0.2173 
   9 |     0.0864      0.2906 

output BN feature 0 mean 3.88578e-17	output BN std 1.05404
output BN feature 1 mean 0	output BN std 1.05396
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    0.08656      0.8613      0.3256      -1.282 
   1 |    -0.2163       1.732       1.159        2.45 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.3399     -0.6616     -0.6168     -0.5781 
   1 |    -0.1755     -0.6403     -0.2351      0.4382 

 training batch 2 mu var00.245256
compute loss for weight  0.339885  0.339875 result 2.16653
 training batch 3 mu var00.245253
compute loss for weight  0.339865  0.339875 result 2.16653
 training batch 4 mu var00.245254
compute loss for weight  0.33988  0.339875 result 2.16653
 training batch 5 mu var00.245253
compute loss for weight  0.33987  0.339875 result 2.16653
   --dy = 0.0865588 dy_ref = 0.0865588
 training batch 6 mu var00.245253
compute loss for weight  -0.661549  -0.661559 result 2.16654
 training batch 7 mu var00.245253
compute loss for weight  -0.661569  -0.661559 result 2.16652
 training batch 8 mu var00.245253
compute loss for weight  -0.661554  -0.661559 result 2.16653
 training batch 9 mu var00.245253
compute loss for weight  -0.661564  -0.661559 result 2.16652
   --dy = 0.861259 dy_ref = 0.861259
 training batch 10 mu var00.245254
compute loss for weight  -0.616815  -0.616825 result 2.16653
 training batch 11 mu var00.245253
compute loss for weight  -0.616835  -0.616825 result 2.16652
 training batch 12 mu var00.245253
compute loss for weight  -0.61682  -0.616825 result 2.16653
 training batch 13 mu var00.245253
compute loss for weight  -0.61683  -0.616825 result 2.16653
   --dy = 0.325632 dy_ref = 0.325632
 training batch 14 mu var00.245253
compute loss for weight  -0.578088  -0.578098 result 2.16651
 training batch 15 mu var00.245253
compute loss for weight  -0.578108  -0.578098 result 2.16654
 training batch 16 mu var00.245253
compute loss for weight  -0.578093  -0.578098 result 2.16652
 training batch 17 mu var00.245253
compute loss for weight  -0.578103  -0.578098 result 2.16653
   --dy = -1.28218 dy_ref = -1.28218
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1667       4.166 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.245253
compute loss for weight  1.00001  1 result 2.16653
 training batch 19 mu var00.245253
compute loss for weight  0.99999  1 result 2.16652
 training batch 20 mu var00.245253
compute loss for weight  1.00001  1 result 2.16653
 training batch 21 mu var00.245253
compute loss for weight  0.999995  1 result 2.16653
   --dy = 0.166686 dy_ref = 0.166686
 training batch 22 mu var00.245253
compute loss for weight  1.00001  1 result 2.16657
 training batch 23 mu var00.245253
compute loss for weight  0.99999  1 result 2.16648
 training batch 24 mu var00.245253
compute loss for weight  1.00001  1 result 2.16655
 training batch 25 mu var00.245253
compute loss for weight  0.999995  1 result 2.16651
   --dy = 4.16637 dy_ref = 4.16637
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  9.021e-17   -1.11e-16 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.245253
compute loss for weight  1e-05  0 result 2.16653
 training batch 27 mu var00.245253
compute loss for weight  -1e-05  0 result 2.16653
 training batch 28 mu var00.245253
compute loss for weight  5e-06  0 result 2.16653
 training batch 29 mu var00.245253
compute loss for weight  -5e-06  0 result 2.16653
   --dy = 0 dy_ref = 9.02056e-17
 training batch 30 mu var00.245253
compute loss for weight  1e-05  0 result 2.16653
 training batch 31 mu var00.245253
compute loss for weight  -1e-05  0 result 2.16653
 training batch 32 mu var00.245253
compute loss for weight  5e-06  0 result 2.16653
 training batch 33 mu var00.245253
compute loss for weight  -5e-06  0 result 2.16653
   --dy = 4.44089e-11 dy_ref = -1.11022e-16
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2308      -2.624 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.7222      -1.588 

 training batch 34 mu var00.245253
compute loss for weight  0.722245  0.722235 result 2.16653
 training batch 35 mu var00.245253
compute loss for weight  0.722225  0.722235 result 2.16652
 training batch 36 mu var00.245253
compute loss for weight  0.72224  0.722235 result 2.16653
 training batch 37 mu var00.245253
compute loss for weight  0.72223  0.722235 result 2.16653
   --dy = 0.230793 dy_ref = 0.230793
 training batch 38 mu var00.245253
compute loss for weight  -1.58808  -1.58809 result 2.1665
 training batch 39 mu var00.245253
compute loss for weight  -1.5881  -1.58809 result 2.16655
 training batch 40 mu var00.245253
compute loss for weight  -1.58809  -1.58809 result 2.16651
 training batch 41 mu var00.245253
compute loss for weight  -1.5881  -1.58809 result 2.16654
   --dy = -2.62351 dy_ref = -2.62351
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m4.7777e-10[NON-XML-CHAR-0x1B][39m