Execution Time0.06s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-ubuntu19-gcc8 (root-ubuntu1904-1) on 2019-11-14 01:35:19
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 var00.599624
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.8291      0.2261 
   1 |      1.274     -0.2131 
   2 |     0.1552      0.2282 
   3 |      2.136      0.5727 
   4 |      1.096      0.3692 
   5 |      2.356      0.5606 
   6 |    -0.8015      0.7483 
   7 |     -1.186      -1.428 
   8 |   -0.06523      0.1362 
   9 |     0.2032     0.09837 

output BN 
output DL feature 0 mean 0.599624	output DL std 1.16174
output DL feature 1 mean 0.129866	output DL std 0.613382
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2082      0.1653 
   1 |     0.6118     -0.5893 
   2 |    -0.4033       0.169 
   3 |      1.394      0.7609 
   4 |     0.4508      0.4113 
   5 |      1.593      0.7402 
   6 |     -1.271       1.063 
   7 |      -1.62      -2.677 
   8 |    -0.6032     0.01093 
   9 |    -0.3596    -0.05412 

output BN feature 0 mean 2.22045e-17	output BN std 1.05405
output BN feature 1 mean 6.38378e-17	output BN std 1.05394
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 | -4.689e-06     0.00265  -0.0008112   -0.002143 
   1 |    -0.2176      0.8846     -0.1065     -0.1466 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.9266     -0.6023     -0.4806     -0.5776 
   1 |    -0.1582     -0.6883       0.117      0.2193 

 training batch 2 mu var00.599627
compute loss for weight  0.926565  0.926555 result 0.16203
 training batch 3 mu var00.599624
compute loss for weight  0.926545  0.926555 result 0.16203
 training batch 4 mu var00.599625
compute loss for weight  0.92656  0.926555 result 0.16203
 training batch 5 mu var00.599624
compute loss for weight  0.92655  0.926555 result 0.16203
   --dy = -4.68916e-06 dy_ref = -4.68916e-06
 training batch 6 mu var00.599623
compute loss for weight  -0.602333  -0.602343 result 0.16203
 training batch 7 mu var00.599624
compute loss for weight  -0.602353  -0.602343 result 0.16203
 training batch 8 mu var00.599624
compute loss for weight  -0.602338  -0.602343 result 0.16203
 training batch 9 mu var00.599624
compute loss for weight  -0.602348  -0.602343 result 0.16203
   --dy = 0.00265005 dy_ref = 0.00265005
 training batch 10 mu var00.599624
compute loss for weight  -0.480572  -0.480582 result 0.16203
 training batch 11 mu var00.599624
compute loss for weight  -0.480592  -0.480582 result 0.16203
 training batch 12 mu var00.599624
compute loss for weight  -0.480577  -0.480582 result 0.16203
 training batch 13 mu var00.599624
compute loss for weight  -0.480587  -0.480582 result 0.16203
   --dy = -0.000811178 dy_ref = -0.000811178
 training batch 14 mu var00.599624
compute loss for weight  -0.577602  -0.577612 result 0.16203
 training batch 15 mu var00.599624
compute loss for weight  -0.577622  -0.577612 result 0.16203
 training batch 16 mu var00.599624
compute loss for weight  -0.577607  -0.577612 result 0.16203
 training batch 17 mu var00.599624
compute loss for weight  -0.577617  -0.577612 result 0.16203
   --dy = -0.00214272 dy_ref = -0.00214272
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      0.327   -0.002944 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.599624
compute loss for weight  1.00001  1 result 0.162034
 training batch 19 mu var00.599624
compute loss for weight  0.99999  1 result 0.162027
 training batch 20 mu var00.599624
compute loss for weight  1.00001  1 result 0.162032
 training batch 21 mu var00.599624
compute loss for weight  0.999995  1 result 0.162029
   --dy = 0.327005 dy_ref = 0.327005
 training batch 22 mu var00.599624
compute loss for weight  1.00001  1 result 0.16203
 training batch 23 mu var00.599624
compute loss for weight  0.99999  1 result 0.16203
 training batch 24 mu var00.599624
compute loss for weight  1.00001  1 result 0.16203
 training batch 25 mu var00.599624
compute loss for weight  0.999995  1 result 0.16203
   --dy = -0.00294427 dy_ref = -0.00294427
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  3.123e-17  -2.439e-19 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.599624
compute loss for weight  1e-05  0 result 0.16203
 training batch 27 mu var00.599624
compute loss for weight  -1e-05  0 result 0.16203
 training batch 28 mu var00.599624
compute loss for weight  5e-06  0 result 0.16203
 training batch 29 mu var00.599624
compute loss for weight  -5e-06  0 result 0.16203
   --dy = 3.23815e-12 dy_ref = 3.1225e-17
 training batch 30 mu var00.599624
compute loss for weight  1e-05  0 result 0.16203
 training batch 31 mu var00.599624
compute loss for weight  -1e-05  0 result 0.16203
 training batch 32 mu var00.599624
compute loss for weight  5e-06  0 result 0.16203
 training batch 33 mu var00.599624
compute loss for weight  -5e-06  0 result 0.16203
   --dy = 0 dy_ref = -2.43945e-19
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.8049     -0.3938 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.4063    0.007477 

 training batch 34 mu var00.599624
compute loss for weight  -0.406246  -0.406256 result 0.162022
 training batch 35 mu var00.599624
compute loss for weight  -0.406266  -0.406256 result 0.162038
 training batch 36 mu var00.599624
compute loss for weight  -0.406251  -0.406256 result 0.162026
 training batch 37 mu var00.599624
compute loss for weight  -0.406261  -0.406256 result 0.162034
   --dy = -0.804923 dy_ref = -0.804923
 training batch 38 mu var00.599624
compute loss for weight  0.00748702  0.00747702 result 0.162026
 training batch 39 mu var00.599624
compute loss for weight  0.00746702  0.00747702 result 0.162034
 training batch 40 mu var00.599624
compute loss for weight  0.00748202  0.00747702 result 0.162028
 training batch 41 mu var00.599624
compute loss for weight  0.00747202  0.00747702 result 0.162032
   --dy = -0.393776 dy_ref = -0.393776
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m1.35587e-06[NON-XML-CHAR-0x1B][39m