Execution Time0.09s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-fedora29-gcc8 (root-fedora29-3.cern.ch) on 2019-11-14 04:22:33

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 var0-0.359179
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.9078       0.136 
   1 |     0.7605      0.9179 
   2 |     -1.996      0.4934 
   3 |      1.361       4.207 
   4 |    -0.6191       1.041 
   5 |     -1.377       1.485 
   6 |      1.975      -1.004 
   7 |     -1.788      -2.407 
   8 |     -1.452      -2.361 
   9 |     0.4526      0.6082 

output BN 
output DL feature 0 mean -0.359179	output DL std 1.4001
output DL feature 1 mean 0.311655	output DL std 1.93969
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -0.413    -0.09543 
   1 |      0.843      0.3294 
   2 |     -1.232     0.09875 
   3 |      1.295       2.117 
   4 |    -0.1957      0.3961 
   5 |    -0.7662      0.6378 
   6 |      1.757     -0.7151 
   7 |     -1.076      -1.477 
   8 |    -0.8229      -1.452 
   9 |     0.6111      0.1612 

output BN feature 0 mean -5.55112e-17	output BN std 1.05406
output BN feature 1 mean 2.77556e-18	output BN std 1.05408
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    0.06239     -0.3388    -0.05308     -0.4225 
   1 |    -0.1187      0.2435    -0.09841      0.4443 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    -0.3247     -0.7797      -1.152       0.722 
   1 |    0.09323      -1.402     -0.7629      -0.993 

 training batch 2 mu var0-0.359176
compute loss for weight  -0.324735  -0.324745 result 1.14494
 training batch 3 mu var0-0.359179
compute loss for weight  -0.324755  -0.324745 result 1.14494
 training batch 4 mu var0-0.359178
compute loss for weight  -0.32474  -0.324745 result 1.14494
 training batch 5 mu var0-0.359179
compute loss for weight  -0.32475  -0.324745 result 1.14494
   --dy = 0.0623935 dy_ref = 0.0623935
 training batch 6 mu var0-0.35918
compute loss for weight  -0.779699  -0.779709 result 1.14494
 training batch 7 mu var0-0.359179
compute loss for weight  -0.779719  -0.779709 result 1.14494
 training batch 8 mu var0-0.359179
compute loss for weight  -0.779704  -0.779709 result 1.14494
 training batch 9 mu var0-0.359179
compute loss for weight  -0.779714  -0.779709 result 1.14494
   --dy = -0.338831 dy_ref = -0.338831
 training batch 10 mu var0-0.359179
compute loss for weight  -1.15209  -1.1521 result 1.14494
 training batch 11 mu var0-0.359179
compute loss for weight  -1.15211  -1.1521 result 1.14494
 training batch 12 mu var0-0.359179
compute loss for weight  -1.1521  -1.1521 result 1.14494
 training batch 13 mu var0-0.359179
compute loss for weight  -1.15211  -1.1521 result 1.14494
   --dy = -0.0530755 dy_ref = -0.0530755
 training batch 14 mu var0-0.359179
compute loss for weight  0.721993  0.721983 result 1.14494
 training batch 15 mu var0-0.359179
compute loss for weight  0.721973  0.721983 result 1.14495
 training batch 16 mu var0-0.359179
compute loss for weight  0.721988  0.721983 result 1.14494
 training batch 17 mu var0-0.359179
compute loss for weight  0.721978  0.721983 result 1.14494
   --dy = -0.422475 dy_ref = -0.422475
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.9955       1.294 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.359179
compute loss for weight  1.00001  1 result 1.14495
 training batch 19 mu var0-0.359179
compute loss for weight  0.99999  1 result 1.14493
 training batch 20 mu var0-0.359179
compute loss for weight  1.00001  1 result 1.14495
 training batch 21 mu var0-0.359179
compute loss for weight  0.999995  1 result 1.14494
   --dy = 0.995491 dy_ref = 0.995491
 training batch 22 mu var0-0.359179
compute loss for weight  1.00001  1 result 1.14495
 training batch 23 mu var0-0.359179
compute loss for weight  0.99999  1 result 1.14493
 training batch 24 mu var0-0.359179
compute loss for weight  1.00001  1 result 1.14495
 training batch 25 mu var0-0.359179
compute loss for weight  0.999995  1 result 1.14493
   --dy = 1.29439 dy_ref = 1.29439
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -9.714e-17  -2.776e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.359179
compute loss for weight  1e-05  0 result 1.14494
 training batch 27 mu var0-0.359179
compute loss for weight  -1e-05  0 result 1.14494
 training batch 28 mu var0-0.359179
compute loss for weight  5e-06  0 result 1.14494
 training batch 29 mu var0-0.359179
compute loss for weight  -5e-06  0 result 1.14494
   --dy = 3.70074e-12 dy_ref = -9.71445e-17
 training batch 30 mu var0-0.359179
compute loss for weight  1e-05  0 result 1.14494
 training batch 31 mu var0-0.359179
compute loss for weight  -1e-05  0 result 1.14494
 training batch 32 mu var0-0.359179
compute loss for weight  5e-06  0 result 1.14494
 training batch 33 mu var0-0.359179
compute loss for weight  -5e-06  0 result 1.14494
   --dy = -3.33067e-11 dy_ref = -2.77556e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.717      -1.857 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.5799     -0.6969 

 training batch 34 mu var0-0.359179
compute loss for weight  -0.57986  -0.57987 result 1.14492
 training batch 35 mu var0-0.359179
compute loss for weight  -0.57988  -0.57987 result 1.14496
 training batch 36 mu var0-0.359179
compute loss for weight  -0.579865  -0.57987 result 1.14493
 training batch 37 mu var0-0.359179
compute loss for weight  -0.579875  -0.57987 result 1.14495
   --dy = -1.71675 dy_ref = -1.71675
 training batch 38 mu var0-0.359179
compute loss for weight  -0.696909  -0.696919 result 1.14492
 training batch 39 mu var0-0.359179
compute loss for weight  -0.696929  -0.696919 result 1.14496
 training batch 40 mu var0-0.359179
compute loss for weight  -0.696914  -0.696919 result 1.14493
 training batch 41 mu var0-0.359179
compute loss for weight  -0.696924  -0.696919 result 1.14495
   --dy = -1.85731 dy_ref = -1.85731
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m7.88652e-10[NON-XML-CHAR-0x1B][39m