Execution Time1.11s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-ubuntu18-clang91-dbg (sft-ubuntu-1804-3) on 2019-11-15 23:17:03
Repository revision: cdf8874e8c83beaf9bc39a224629667fbb7903bc

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 var00.107498
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1463     0.08921 
   1 |     0.3122       1.102 
   2 |    -0.2835       0.771 
   3 |     0.2803       4.866 
   4 |     0.1505       1.145 
   5 |     0.3941       1.598 
   6 |      0.363      -1.908 
   7 |    -0.5225      -2.061 
   8 |     0.1857      -3.006 
   9 |    0.04881       0.673 

output BN 
output DL feature 0 mean 0.107498	output DL std 0.294743
output DL feature 1 mean 0.326878	output DL std 2.25245
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1387     -0.1112 
   1 |     0.7316      0.3626 
   2 |     -1.397      0.2079 
   3 |     0.6177       2.124 
   4 |     0.1536      0.3828 
   5 |      1.024      0.5948 
   6 |     0.9132      -1.046 
   7 |     -2.252      -1.117 
   8 |     0.2794       -1.56 
   9 |    -0.2098       0.162 

output BN feature 0 mean 3.33067e-17	output BN std 1.05342
output BN feature 1 mean 0	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 |     -0.364     -0.9368      -0.755      -1.834 
   1 |     0.1223     -0.1196    0.004428      0.1877 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.1921     -0.2088     -0.1691      0.1376 
   1 |     0.1556      -1.321     -0.8758      -1.467 

 training batch 2 mu var00.107501
compute loss for weight  0.192063  0.192053 result 0.855339
 training batch 3 mu var00.107498
compute loss for weight  0.192043  0.192053 result 0.855347
 training batch 4 mu var00.107499
compute loss for weight  0.192058  0.192053 result 0.855341
 training batch 5 mu var00.107498
compute loss for weight  0.192048  0.192053 result 0.855345
   --dy = -0.364038 dy_ref = -0.364038
 training batch 6 mu var00.107498
compute loss for weight  -0.208801  -0.208811 result 0.855334
 training batch 7 mu var00.107498
compute loss for weight  -0.208821  -0.208811 result 0.855352
 training batch 8 mu var00.107498
compute loss for weight  -0.208806  -0.208811 result 0.855338
 training batch 9 mu var00.107498
compute loss for weight  -0.208816  -0.208811 result 0.855348
   --dy = -0.936764 dy_ref = -0.936764
 training batch 10 mu var00.107499
compute loss for weight  -0.169071  -0.169081 result 0.855335
 training batch 11 mu var00.107498
compute loss for weight  -0.169091  -0.169081 result 0.85535
 training batch 12 mu var00.107499
compute loss for weight  -0.169076  -0.169081 result 0.855339
 training batch 13 mu var00.107498
compute loss for weight  -0.169086  -0.169081 result 0.855347
   --dy = -0.755032 dy_ref = -0.755032
 training batch 14 mu var00.107498
compute loss for weight  0.137583  0.137573 result 0.855325
 training batch 15 mu var00.107498
compute loss for weight  0.137563  0.137573 result 0.855361
 training batch 16 mu var00.107498
compute loss for weight  0.137578  0.137573 result 0.855334
 training batch 17 mu var00.107498
compute loss for weight  0.137568  0.137573 result 0.855352
   --dy = -1.83435 dy_ref = -1.83435
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.7809      0.9298 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.107498
compute loss for weight  1.00001  1 result 0.855351
 training batch 19 mu var00.107498
compute loss for weight  0.99999  1 result 0.855335
 training batch 20 mu var00.107498
compute loss for weight  1.00001  1 result 0.855347
 training batch 21 mu var00.107498
compute loss for weight  0.999995  1 result 0.855339
   --dy = 0.780925 dy_ref = 0.780925
 training batch 22 mu var00.107498
compute loss for weight  1.00001  1 result 0.855352
 training batch 23 mu var00.107498
compute loss for weight  0.99999  1 result 0.855334
 training batch 24 mu var00.107498
compute loss for weight  1.00001  1 result 0.855348
 training batch 25 mu var00.107498
compute loss for weight  0.999995  1 result 0.855338
   --dy = 0.929761 dy_ref = 0.929761
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -1.084e-17  -2.927e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.107498
compute loss for weight  1e-05  0 result 0.855343
 training batch 27 mu var00.107498
compute loss for weight  -1e-05  0 result 0.855343
 training batch 28 mu var00.107498
compute loss for weight  5e-06  0 result 0.855343
 training batch 29 mu var00.107498
compute loss for weight  -5e-06  0 result 0.855343
   --dy = 3.70074e-12 dy_ref = -1.0842e-17
 training batch 30 mu var00.107498
compute loss for weight  1e-05  0 result 0.855343
 training batch 31 mu var00.107498
compute loss for weight  -1e-05  0 result 0.855343
 training batch 32 mu var00.107498
compute loss for weight  5e-06  0 result 0.855343
 training batch 33 mu var00.107498
compute loss for weight  -5e-06  0 result 0.855343
   --dy = -2.77556e-11 dy_ref = -2.92735e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.446      -1.537 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.5402     -0.6049 

 training batch 34 mu var00.107498
compute loss for weight  -0.540184  -0.540194 result 0.855328
 training batch 35 mu var00.107498
compute loss for weight  -0.540204  -0.540194 result 0.855357
 training batch 36 mu var00.107498
compute loss for weight  -0.540189  -0.540194 result 0.855336
 training batch 37 mu var00.107498
compute loss for weight  -0.540199  -0.540194 result 0.85535
   --dy = -1.44564 dy_ref = -1.44564
 training batch 38 mu var00.107498
compute loss for weight  -0.604856  -0.604866 result 0.855328
 training batch 39 mu var00.107498
compute loss for weight  -0.604876  -0.604866 result 0.855358
 training batch 40 mu var00.107498
compute loss for weight  -0.604861  -0.604866 result 0.855335
 training batch 41 mu var00.107498
compute loss for weight  -0.604871  -0.604866 result 0.855351
   --dy = -1.53713 dy_ref = -1.53713
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m2.77555e-11[NON-XML-CHAR-0x1B][39m