Execution Time0.09s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-centos7-gcc48-dbg (lcgapp-centos7-x86-64-25.cern.ch) on 2019-11-21 08:37:58

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 var00.374401
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.7482     -0.8571 
   1 |    -0.8963      0.2165 
   2 |      2.224     -0.6506 
   3 |      1.009        2.09 
   4 |      1.074     -0.2934 
   5 |      1.652      -1.131 
   6 |     -1.216      0.4913 
   7 |    -0.8464       -1.43 
   8 |    0.01197      -2.268 
   9 |   -0.01618      0.4746 

output BN 
output DL feature 0 mean 0.374401	output DL std 1.15482
output DL feature 1 mean -0.335847	output DL std 1.22326
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.3412     -0.4491 
   1 |      -1.16       0.476 
   2 |      1.688     -0.2712 
   3 |     0.5792        2.09 
   4 |     0.6389     0.03659 
   5 |      1.166     -0.6855 
   6 |     -1.452      0.7127 
   7 |     -1.114     -0.9428 
   8 |    -0.3308      -1.665 
   9 |    -0.3565      0.6984 

output BN feature 0 mean -4.996e-17	output BN std 1.05405
output BN feature 1 mean -3.33067e-17	output BN std 1.05405
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |      1.027       1.202       1.765      0.7717 
   1 |   -0.04386        1.77     -0.2508        1.74 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    -0.2324     -0.6139      0.8836     -0.7553 
   1 |    -0.6216      -0.805     -0.7525    -0.07158 

 training batch 2 mu var00.374404
compute loss for weight  -0.23237  -0.23238 result 3.84513
 training batch 3 mu var00.374401
compute loss for weight  -0.23239  -0.23238 result 3.84511
 training batch 4 mu var00.374402
compute loss for weight  -0.232375  -0.23238 result 3.84512
 training batch 5 mu var00.374401
compute loss for weight  -0.232385  -0.23238 result 3.84511
   --dy = 1.02691 dy_ref = 1.02691
 training batch 6 mu var00.374401
compute loss for weight  -0.613938  -0.613948 result 3.84513
 training batch 7 mu var00.374401
compute loss for weight  -0.613958  -0.613948 result 3.84511
 training batch 8 mu var00.374401
compute loss for weight  -0.613943  -0.613948 result 3.84512
 training batch 9 mu var00.374401
compute loss for weight  -0.613953  -0.613948 result 3.84511
   --dy = 1.20175 dy_ref = 1.20175
 training batch 10 mu var00.374402
compute loss for weight  0.883562  0.883552 result 3.84514
 training batch 11 mu var00.374401
compute loss for weight  0.883542  0.883552 result 3.8451
 training batch 12 mu var00.374402
compute loss for weight  0.883557  0.883552 result 3.84513
 training batch 13 mu var00.374401
compute loss for weight  0.883547  0.883552 result 3.84511
   --dy = 1.76492 dy_ref = 1.76492
 training batch 14 mu var00.374401
compute loss for weight  -0.755292  -0.755302 result 3.84513
 training batch 15 mu var00.374401
compute loss for weight  -0.755312  -0.755302 result 3.84511
 training batch 16 mu var00.374401
compute loss for weight  -0.755297  -0.755302 result 3.84512
 training batch 17 mu var00.374401
compute loss for weight  -0.755307  -0.755302 result 3.84511
   --dy = 0.771666 dy_ref = 0.771666
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.413       6.277 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.374401
compute loss for weight  1.00001  1 result 3.84513
 training batch 19 mu var00.374401
compute loss for weight  0.99999  1 result 3.8451
 training batch 20 mu var00.374401
compute loss for weight  1.00001  1 result 3.84513
 training batch 21 mu var00.374401
compute loss for weight  0.999995  1 result 3.84511
   --dy = 1.41288 dy_ref = 1.41288
 training batch 22 mu var00.374401
compute loss for weight  1.00001  1 result 3.84518
 training batch 23 mu var00.374401
compute loss for weight  0.99999  1 result 3.84506
 training batch 24 mu var00.374401
compute loss for weight  1.00001  1 result 3.84515
 training batch 25 mu var00.374401
compute loss for weight  0.999995  1 result 3.84509
   --dy = 6.27736 dy_ref = 6.27736
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  1.943e-16  -4.441e-16 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.374401
compute loss for weight  1e-05  0 result 3.84512
 training batch 27 mu var00.374401
compute loss for weight  -1e-05  0 result 3.84512
 training batch 28 mu var00.374401
compute loss for weight  5e-06  0 result 3.84512
 training batch 29 mu var00.374401
compute loss for weight  -5e-06  0 result 3.84512
   --dy = -7.40149e-12 dy_ref = 1.94289e-16
 training batch 30 mu var00.374401
compute loss for weight  1e-05  0 result 3.84512
 training batch 31 mu var00.374401
compute loss for weight  -1e-05  0 result 3.84512
 training batch 32 mu var00.374401
compute loss for weight  5e-06  0 result 3.84512
 training batch 33 mu var00.374401
compute loss for weight  -5e-06  0 result 3.84512
   --dy = 1.4803e-11 dy_ref = -4.44089e-16
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.755       3.576 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.8053       1.755 

 training batch 34 mu var00.374401
compute loss for weight  -0.805255  -0.805265 result 3.8451
 training batch 35 mu var00.374401
compute loss for weight  -0.805275  -0.805265 result 3.84514
 training batch 36 mu var00.374401
compute loss for weight  -0.80526  -0.805265 result 3.84511
 training batch 37 mu var00.374401
compute loss for weight  -0.80527  -0.805265 result 3.84513
   --dy = -1.75455 dy_ref = -1.75455
 training batch 38 mu var00.374401
compute loss for weight  1.75525  1.75524 result 3.84515
 training batch 39 mu var00.374401
compute loss for weight  1.75523  1.75524 result 3.84508
 training batch 40 mu var00.374401
compute loss for weight  1.75525  1.75524 result 3.84514
 training batch 41 mu var00.374401
compute loss for weight  1.75524  1.75524 result 3.8451
   --dy = 3.57635 dy_ref = 3.57635
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m2.06834e-10[NON-XML-CHAR-0x1B][39m