Execution Time0.18s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-centos7-gcc48-dbg (olhswep22.cern.ch) on 2019-11-12 13:18:05

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 var00.340095
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.8342     -0.2365 
   1 |    -0.6425       1.546 
   2 |      1.778       -2.15 
   3 |     -1.105       1.416 
   4 |     0.5958     -0.1145 
   5 |      1.309      0.0603 
   6 |     -1.737       1.737 
   7 |      1.477      -1.787 
   8 |      1.278     -0.4669 
   9 |    -0.3878      0.3702 

output BN 
output DL feature 0 mean 0.340095	output DL std 1.21953
output DL feature 1 mean 0.0375875	output DL std 1.31829
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.4271     -0.2192 
   1 |    -0.8493       1.206 
   2 |      1.243      -1.749 
   3 |     -1.249       1.103 
   4 |      0.221     -0.1216 
   5 |     0.8375     0.01816 
   6 |     -1.795       1.359 
   7 |     0.9829      -1.459 
   8 |     0.8109     -0.4033 
   9 |    -0.6291      0.2659 

output BN feature 0 mean -1.11022e-17	output BN std 1.05405
output BN feature 1 mean -1.11022e-17	output BN std 1.05406
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |    0.01474   -0.008768    0.002453    0.002135 
   1 |    0.01523    -0.01338   -0.003285    0.004899 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.3056      0.6378       1.006     -0.6471 
   1 |      0.486     -0.6924      -1.283      0.6628 

 training batch 2 mu var00.340097
compute loss for weight  0.305629  0.305619 result 0.0241721
 training batch 3 mu var00.340095
compute loss for weight  0.305609  0.305619 result 0.0241718
 training batch 4 mu var00.340095
compute loss for weight  0.305624  0.305619 result 0.024172
 training batch 5 mu var00.340095
compute loss for weight  0.305614  0.305619 result 0.0241719
   --dy = 0.014735 dy_ref = 0.014735
 training batch 6 mu var00.340094
compute loss for weight  0.637761  0.637751 result 0.0241719
 training batch 7 mu var00.340095
compute loss for weight  0.637741  0.637751 result 0.024172
 training batch 8 mu var00.340094
compute loss for weight  0.637756  0.637751 result 0.0241719
 training batch 9 mu var00.340095
compute loss for weight  0.637746  0.637751 result 0.024172
   --dy = -0.00876778 dy_ref = -0.00876778
 training batch 10 mu var00.340095
compute loss for weight  1.00638  1.00637 result 0.024172
 training batch 11 mu var00.340095
compute loss for weight  1.00636  1.00637 result 0.0241719
 training batch 12 mu var00.340095
compute loss for weight  1.00638  1.00637 result 0.024172
 training batch 13 mu var00.340095
compute loss for weight  1.00637  1.00637 result 0.0241719
   --dy = 0.00245263 dy_ref = 0.00245263
 training batch 14 mu var00.340095
compute loss for weight  -0.647094  -0.647104 result 0.024172
 training batch 15 mu var00.340095
compute loss for weight  -0.647114  -0.647104 result 0.0241719
 training batch 16 mu var00.340095
compute loss for weight  -0.647099  -0.647104 result 0.0241719
 training batch 17 mu var00.340095
compute loss for weight  -0.647109  -0.647104 result 0.0241719
   --dy = 0.00213542 dy_ref = 0.00213542
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |   -0.02608     0.07442 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.340095
compute loss for weight  1.00001  1 result 0.0241717
 training batch 19 mu var00.340095
compute loss for weight  0.99999  1 result 0.0241722
 training batch 20 mu var00.340095
compute loss for weight  1.00001  1 result 0.0241718
 training batch 21 mu var00.340095
compute loss for weight  0.999995  1 result 0.0241721
   --dy = -0.0260779 dy_ref = -0.0260779
 training batch 22 mu var00.340095
compute loss for weight  1.00001  1 result 0.0241727
 training batch 23 mu var00.340095
compute loss for weight  0.99999  1 result 0.0241712
 training batch 24 mu var00.340095
compute loss for weight  1.00001  1 result 0.0241723
 training batch 25 mu var00.340095
compute loss for weight  0.999995  1 result 0.0241716
   --dy = 0.0744218 dy_ref = 0.0744218
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -3.063e-18  -3.686e-18 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.340095
compute loss for weight  1e-05  0 result 0.0241719
 training batch 27 mu var00.340095
compute loss for weight  -1e-05  0 result 0.0241719
 training batch 28 mu var00.340095
compute loss for weight  5e-06  0 result 0.0241719
 training batch 29 mu var00.340095
compute loss for weight  -5e-06  0 result 0.0241719
   --dy = 8.67362e-13 dy_ref = -3.06287e-18
 training batch 30 mu var00.340095
compute loss for weight  1e-05  0 result 0.0241719
 training batch 31 mu var00.340095
compute loss for weight  -1e-05  0 result 0.0241719
 training batch 32 mu var00.340095
compute loss for weight  5e-06  0 result 0.0241719
 training batch 33 mu var00.340095
compute loss for weight  -5e-06  0 result 0.0241719
   --dy = 5.78241e-14 dy_ref = -3.68629e-18
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2246     -0.2948 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |    -0.1161     -0.2525 

 training batch 34 mu var00.340095
compute loss for weight  -0.116114  -0.116124 result 0.0241742
 training batch 35 mu var00.340095
compute loss for weight  -0.116134  -0.116124 result 0.0241697
 training batch 36 mu var00.340095
compute loss for weight  -0.116119  -0.116124 result 0.0241731
 training batch 37 mu var00.340095
compute loss for weight  -0.116129  -0.116124 result 0.0241708
   --dy = 0.22457 dy_ref = 0.22457
 training batch 38 mu var00.340095
compute loss for weight  -0.252453  -0.252463 result 0.024169
 training batch 39 mu var00.340095
compute loss for weight  -0.252473  -0.252463 result 0.0241749
 training batch 40 mu var00.340095
compute loss for weight  -0.252458  -0.252463 result 0.0241705
 training batch 41 mu var00.340095
compute loss for weight  -0.252468  -0.252463 result 0.0241734
   --dy = -0.294783 dy_ref = -0.294783
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m6.67759e-10[NON-XML-CHAR-0x1B][39m