Execution Time0.61s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-mac1013-clang100 (macphsft16.dyndns.cern.ch) on 2019-11-14 00:49:58

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 var00.273203
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.4877      0.8935 
   1 |    -0.2942     -0.7909 
   2 |      2.361      0.6659 
   3 |     0.9919      -3.146 
   4 |     0.7735      0.1394 
   5 |      1.236      0.8482 
   6 |     -3.649      0.4374 
   7 |      1.785       1.222 
   8 |    -0.8537       3.072 
   9 |    -0.1071     -0.6049 

output BN 
output DL feature 0 mean 0.273203	output DL std 1.6834
output DL feature 1 mean 0.273711	output DL std 1.60708
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.1343      0.4065 
   1 |    -0.3553     -0.6982 
   2 |      1.308      0.2572 
   3 |       0.45      -2.243 
   4 |     0.3132    -0.08812 
   5 |      0.603      0.3768 
   6 |     -2.456      0.1074 
   7 |     0.9466      0.6221 
   8 |    -0.7056       1.836 
   9 |    -0.2381     -0.5763 

output BN feature 0 mean 0	output BN std 1.05407
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 |    0.06399     0.07768       0.135     0.07824 
   1 |    -0.2786     -0.3253     -0.5504     -0.5783 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.1921      0.5445      0.6137      -1.756 
   1 |     0.3732      0.7787       1.122      0.5864 

 training batch 2 mu var00.273205
compute loss for weight  0.192069  0.192059 result 0.6468
 training batch 3 mu var00.273203
compute loss for weight  0.192049  0.192059 result 0.646799
 training batch 4 mu var00.273203
compute loss for weight  0.192064  0.192059 result 0.6468
 training batch 5 mu var00.273203
compute loss for weight  0.192054  0.192059 result 0.646799
   --dy = 0.0639852 dy_ref = 0.0639852
 training batch 6 mu var00.273202
compute loss for weight  0.544488  0.544478 result 0.646801
 training batch 7 mu var00.273203
compute loss for weight  0.544468  0.544478 result 0.646799
 training batch 8 mu var00.273202
compute loss for weight  0.544483  0.544478 result 0.6468
 training batch 9 mu var00.273203
compute loss for weight  0.544473  0.544478 result 0.646799
   --dy = 0.0776796 dy_ref = 0.0776796
 training batch 10 mu var00.273203
compute loss for weight  0.613676  0.613666 result 0.646801
 training batch 11 mu var00.273203
compute loss for weight  0.613656  0.613666 result 0.646798
 training batch 12 mu var00.273203
compute loss for weight  0.613671  0.613666 result 0.6468
 training batch 13 mu var00.273203
compute loss for weight  0.613661  0.613666 result 0.646799
   --dy = 0.135022 dy_ref = 0.135022
 training batch 14 mu var00.273203
compute loss for weight  -1.75612  -1.75613 result 0.646801
 training batch 15 mu var00.273203
compute loss for weight  -1.75614  -1.75613 result 0.646799
 training batch 16 mu var00.273203
compute loss for weight  -1.75612  -1.75613 result 0.6468
 training batch 17 mu var00.273203
compute loss for weight  -1.75613  -1.75613 result 0.646799
   --dy = 0.0782361 dy_ref = 0.0782361
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.254     0.03927 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.273203
compute loss for weight  1.00001  1 result 0.646812
 training batch 19 mu var00.273203
compute loss for weight  0.99999  1 result 0.646787
 training batch 20 mu var00.273203
compute loss for weight  1.00001  1 result 0.646806
 training batch 21 mu var00.273203
compute loss for weight  0.999995  1 result 0.646794
   --dy = 1.25433 dy_ref = 1.25433
 training batch 22 mu var00.273203
compute loss for weight  1.00001  1 result 0.6468
 training batch 23 mu var00.273203
compute loss for weight  0.99999  1 result 0.646799
 training batch 24 mu var00.273203
compute loss for weight  1.00001  1 result 0.6468
 training batch 25 mu var00.273203
compute loss for weight  0.999995  1 result 0.6468
   --dy = 0.039274 dy_ref = 0.039274
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |   1.18e-16   1.735e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.273203
compute loss for weight  1e-05  0 result 0.6468
 training batch 27 mu var00.273203
compute loss for weight  -1e-05  0 result 0.6468
 training batch 28 mu var00.273203
compute loss for weight  5e-06  0 result 0.6468
 training batch 29 mu var00.273203
compute loss for weight  -5e-06  0 result 0.6468
   --dy = 3.33067e-11 dy_ref = 1.17961e-16
 training batch 30 mu var00.273203
compute loss for weight  1e-05  0 result 0.6468
 training batch 31 mu var00.273203
compute loss for weight  -1e-05  0 result 0.6468
 training batch 32 mu var00.273203
compute loss for weight  5e-06  0 result 0.6468
 training batch 33 mu var00.273203
compute loss for weight  -5e-06  0 result 0.6468
   --dy = -1.29526e-11 dy_ref = 1.73472e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.565      0.2112 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.8013       0.186 

 training batch 34 mu var00.273203
compute loss for weight  0.801348  0.801338 result 0.646815
 training batch 35 mu var00.273203
compute loss for weight  0.801328  0.801338 result 0.646784
 training batch 36 mu var00.273203
compute loss for weight  0.801343  0.801338 result 0.646808
 training batch 37 mu var00.273203
compute loss for weight  0.801333  0.801338 result 0.646792
   --dy = 1.56529 dy_ref = 1.56529
 training batch 38 mu var00.273203
compute loss for weight  0.186004  0.185994 result 0.646802
 training batch 39 mu var00.273203
compute loss for weight  0.185984  0.185994 result 0.646798
 training batch 40 mu var00.273203
compute loss for weight  0.185999  0.185994 result 0.646801
 training batch 41 mu var00.273203
compute loss for weight  0.185989  0.185994 result 0.646799
   --dy = 0.211157 dy_ref = 0.211157
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m2.94736e-10[NON-XML-CHAR-0x1B][39m