Execution Time0.13s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-centos7-clang100-opt-no-rt-cxxmodules (olsnba08.cern.ch) on 2019-11-14 01:39:53
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 var0-0.471439
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -0.884      -2.033 
   1 |    -0.8098       -1.44 
   2 |      1.062      0.2288 
   3 |     0.8281     -0.1856 
   4 |    -0.4481      -1.651 
   5 |     -1.601      -4.221 
   6 |     -1.931       -1.11 
   7 |      1.284       2.058 
   8 |     -2.348      -3.275 
   9 |      0.134      0.1609 

output BN 
output DL feature 0 mean -0.471439	output DL std 1.27812
output DL feature 1 mean -1.14671	output DL std 1.82271
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.3402     -0.5123 
   1 |     -0.279     -0.1694 
   2 |      1.265      0.7955 
   3 |      1.072      0.5558 
   4 |    0.01926     -0.2919 
   5 |    -0.9319      -1.778 
   6 |     -1.204      0.0215 
   7 |      1.448       1.853 
   8 |     -1.547      -1.231 
   9 |     0.4993      0.7562 

output BN feature 0 mean -2.77556e-17	output BN std 1.05406
output BN feature 1 mean 1.11022e-17	output BN std 1.05407
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |  -0.002073     0.00219   -0.001357     0.00224 
   1 |     0.3546     -0.2502      0.2057      0.1459 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     -0.891      0.3206      0.1472      -1.011 
   1 |     -1.745      0.7044    0.001326     -0.5382 

 training batch 2 mu var0-0.471437
compute loss for weight  -0.891027  -0.891037 result 0.626754
 training batch 3 mu var0-0.471439
compute loss for weight  -0.891047  -0.891037 result 0.626754
 training batch 4 mu var0-0.471439
compute loss for weight  -0.891032  -0.891037 result 0.626754
 training batch 5 mu var0-0.471439
compute loss for weight  -0.891042  -0.891037 result 0.626754
   --dy = -0.00207274 dy_ref = -0.00207274
 training batch 6 mu var0-0.47144
compute loss for weight  0.320588  0.320578 result 0.626754
 training batch 7 mu var0-0.471439
compute loss for weight  0.320568  0.320578 result 0.626754
 training batch 8 mu var0-0.47144
compute loss for weight  0.320583  0.320578 result 0.626754
 training batch 9 mu var0-0.471439
compute loss for weight  0.320573  0.320578 result 0.626754
   --dy = 0.00218962 dy_ref = 0.00218962
 training batch 10 mu var0-0.471439
compute loss for weight  0.147245  0.147235 result 0.626754
 training batch 11 mu var0-0.471439
compute loss for weight  0.147225  0.147235 result 0.626754
 training batch 12 mu var0-0.471439
compute loss for weight  0.14724  0.147235 result 0.626754
 training batch 13 mu var0-0.471439
compute loss for weight  0.14723  0.147235 result 0.626754
   --dy = -0.00135701 dy_ref = -0.00135701
 training batch 14 mu var0-0.47144
compute loss for weight  -1.01086  -1.01087 result 0.626754
 training batch 15 mu var0-0.471439
compute loss for weight  -1.01088  -1.01087 result 0.626754
 training batch 16 mu var0-0.471439
compute loss for weight  -1.01087  -1.01087 result 0.626754
 training batch 17 mu var0-0.471439
compute loss for weight  -1.01088  -1.01087 result 0.626754
   --dy = 0.00223995 dy_ref = 0.00223995
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.246    0.007848 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.471439
compute loss for weight  1.00001  1 result 0.626767
 training batch 19 mu var0-0.471439
compute loss for weight  0.99999  1 result 0.626742
 training batch 20 mu var0-0.471439
compute loss for weight  1.00001  1 result 0.626761
 training batch 21 mu var0-0.471439
compute loss for weight  0.999995  1 result 0.626748
   --dy = 1.24566 dy_ref = 1.24566
 training batch 22 mu var0-0.471439
compute loss for weight  1.00001  1 result 0.626755
 training batch 23 mu var0-0.471439
compute loss for weight  0.99999  1 result 0.626754
 training batch 24 mu var0-0.471439
compute loss for weight  1.00001  1 result 0.626754
 training batch 25 mu var0-0.471439
compute loss for weight  0.999995  1 result 0.626754
   --dy = 0.00784752 dy_ref = 0.00784752
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |  4.163e-17   2.168e-19 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.471439
compute loss for weight  1e-05  0 result 0.626754
 training batch 27 mu var0-0.471439
compute loss for weight  -1e-05  0 result 0.626754
 training batch 28 mu var0-0.471439
compute loss for weight  5e-06  0 result 0.626754
 training batch 29 mu var0-0.471439
compute loss for weight  -5e-06  0 result 0.626754
   --dy = 0 dy_ref = 4.16334e-17
 training batch 30 mu var0-0.471439
compute loss for weight  1e-05  0 result 0.626754
 training batch 31 mu var0-0.471439
compute loss for weight  -1e-05  0 result 0.626754
 training batch 32 mu var0-0.471439
compute loss for weight  5e-06  0 result 0.626754
 training batch 33 mu var0-0.471439
compute loss for weight  -5e-06  0 result 0.626754
   --dy = -1.4803e-11 dy_ref = 2.1684e-19
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.583       1.336 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.7868    0.005876 

 training batch 34 mu var0-0.471439
compute loss for weight  0.786765  0.786755 result 0.62677
 training batch 35 mu var0-0.471439
compute loss for weight  0.786745  0.786755 result 0.626739
 training batch 36 mu var0-0.471439
compute loss for weight  0.78676  0.786755 result 0.626762
 training batch 37 mu var0-0.471439
compute loss for weight  0.78675  0.786755 result 0.626747
   --dy = 1.58329 dy_ref = 1.58329
 training batch 38 mu var0-0.471439
compute loss for weight  0.00588574  0.00587574 result 0.626768
 training batch 39 mu var0-0.471439
compute loss for weight  0.00586574  0.00587574 result 0.626741
 training batch 40 mu var0-0.471439
compute loss for weight  0.00588074  0.00587574 result 0.626761
 training batch 41 mu var0-0.471439
compute loss for weight  0.00587074  0.00587574 result 0.626748
   --dy = 1.33558 dy_ref = 1.33558
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m1.6703e-08[NON-XML-CHAR-0x1B][39m