Execution Time0.78s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-ubuntu18-clang91-opt (sft-ubuntu-1804-3) on 2019-11-15 00:31:16
Repository revision: 14de58de35eff907054671888ccc2de0f7f27e77

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 var0-0.137629
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -0.591       1.085 
   1 |     0.9384     -0.2525 
   2 |    -0.5984       1.297 
   3 |      2.498     -0.8157 
   4 |   -0.04578      0.8252 
   5 |    -0.3976       1.915 
   6 |    -0.7288     -0.8591 
   7 |    -0.5833      0.4537 
   8 |     -2.316       1.748 
   9 |     0.4483     -0.3228 

output BN 
output DL feature 0 mean -0.137629	output DL std 1.25766
output DL feature 1 mean 0.507403	output DL std 1.02557
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      -0.38      0.5939 
   1 |     0.9018      -0.781 
   2 |    -0.3862      0.8114 
   3 |      2.209       -1.36 
   4 |    0.07698      0.3266 
   5 |    -0.2179       1.446 
   6 |    -0.4954      -1.404 
   7 |    -0.3735    -0.05523 
   8 |     -1.826       1.276 
   9 |      0.491     -0.8532 

output BN feature 0 mean 1.05471e-16	output BN std 1.05406
output BN feature 1 mean -5.55112e-17	output BN std 1.05404
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.2003       -0.11      0.1884     -0.2363 
   1 |    -0.1093      -0.229      -0.352     -0.3916 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |   -0.02226      -0.421      -0.985     -0.6083 
   1 |     0.5554      0.2329      0.7972     -0.2833 

 training batch 2 mu var0-0.137626
compute loss for weight  -0.0222468  -0.0222568 result 0.224111
 training batch 3 mu var0-0.137629
compute loss for weight  -0.0222668  -0.0222568 result 0.224107
 training batch 4 mu var0-0.137628
compute loss for weight  -0.0222518  -0.0222568 result 0.22411
 training batch 5 mu var0-0.137629
compute loss for weight  -0.0222618  -0.0222568 result 0.224108
   --dy = 0.200308 dy_ref = 0.200308
 training batch 6 mu var0-0.137629
compute loss for weight  -0.421017  -0.421027 result 0.224108
 training batch 7 mu var0-0.137629
compute loss for weight  -0.421037  -0.421027 result 0.22411
 training batch 8 mu var0-0.137629
compute loss for weight  -0.421022  -0.421027 result 0.224108
 training batch 9 mu var0-0.137629
compute loss for weight  -0.421032  -0.421027 result 0.224109
   --dy = -0.11 dy_ref = -0.11
 training batch 10 mu var0-0.137629
compute loss for weight  -0.985004  -0.985014 result 0.224111
 training batch 11 mu var0-0.137629
compute loss for weight  -0.985024  -0.985014 result 0.224107
 training batch 12 mu var0-0.137629
compute loss for weight  -0.985009  -0.985014 result 0.22411
 training batch 13 mu var0-0.137629
compute loss for weight  -0.985019  -0.985014 result 0.224108
   --dy = 0.188383 dy_ref = 0.188383
 training batch 14 mu var0-0.137629
compute loss for weight  -0.608259  -0.608269 result 0.224106
 training batch 15 mu var0-0.137629
compute loss for weight  -0.608279  -0.608269 result 0.224111
 training batch 16 mu var0-0.137629
compute loss for weight  -0.608264  -0.608269 result 0.224107
 training batch 17 mu var0-0.137629
compute loss for weight  -0.608274  -0.608269 result 0.22411
   --dy = -0.236279 dy_ref = -0.236279
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2314      0.2168 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var0-0.137629
compute loss for weight  1.00001  1 result 0.224111
 training batch 19 mu var0-0.137629
compute loss for weight  0.99999  1 result 0.224106
 training batch 20 mu var0-0.137629
compute loss for weight  1.00001  1 result 0.22411
 training batch 21 mu var0-0.137629
compute loss for weight  0.999995  1 result 0.224107
   --dy = 0.231418 dy_ref = 0.231418
 training batch 22 mu var0-0.137629
compute loss for weight  1.00001  1 result 0.224111
 training batch 23 mu var0-0.137629
compute loss for weight  0.99999  1 result 0.224106
 training batch 24 mu var0-0.137629
compute loss for weight  1.00001  1 result 0.22411
 training batch 25 mu var0-0.137629
compute loss for weight  0.999995  1 result 0.224108
   --dy = 0.216799 dy_ref = 0.216799
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |          0  -1.388e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var0-0.137629
compute loss for weight  1e-05  0 result 0.224109
 training batch 27 mu var0-0.137629
compute loss for weight  -1e-05  0 result 0.224109
 training batch 28 mu var0-0.137629
compute loss for weight  5e-06  0 result 0.224109
 training batch 29 mu var0-0.137629
compute loss for weight  -5e-06  0 result 0.224109
   --dy = 1.85037e-12 dy_ref = 0
 training batch 30 mu var0-0.137629
compute loss for weight  1e-05  0 result 0.224109
 training batch 31 mu var0-0.137629
compute loss for weight  -1e-05  0 result 0.224109
 training batch 32 mu var0-0.137629
compute loss for weight  5e-06  0 result 0.224109
 training batch 33 mu var0-0.137629
compute loss for weight  -5e-06  0 result 0.224109
   --dy = -9.25186e-13 dy_ref = -1.38778e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.4027     -0.3815 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.5747     -0.5683 

 training batch 34 mu var0-0.137629
compute loss for weight  -0.574647  -0.574657 result 0.224105
 training batch 35 mu var0-0.137629
compute loss for weight  -0.574667  -0.574657 result 0.224113
 training batch 36 mu var0-0.137629
compute loss for weight  -0.574652  -0.574657 result 0.224107
 training batch 37 mu var0-0.137629
compute loss for weight  -0.574662  -0.574657 result 0.224111
   --dy = -0.402707 dy_ref = -0.402707
 training batch 38 mu var0-0.137629
compute loss for weight  -0.568261  -0.568271 result 0.224105
 training batch 39 mu var0-0.137629
compute loss for weight  -0.568281  -0.568271 result 0.224112
 training batch 40 mu var0-0.137629
compute loss for weight  -0.568266  -0.568271 result 0.224107
 training batch 41 mu var0-0.137629
compute loss for weight  -0.568276  -0.568271 result 0.224111
   --dy = -0.381506 dy_ref = -0.381506
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m4.38795e-11[NON-XML-CHAR-0x1B][39m