Execution Time0.48s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-mac1015-clang110 (macphsft18.dyndns.cern.ch) on 2019-11-14 00:48:39

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 var00.015445
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2933       1.047 
   1 |    -0.5504     -0.9883 
   2 |      0.526       1.574 
   3 |     -2.291      -1.765 
   4 |    -0.2142      0.6589 
   5 |   -0.07349       1.498 
   6 |    -0.5763      -0.267 
   7 |      1.968      0.4277 
   8 |      1.514       2.274 
   9 |    -0.4409     -0.4464 

output BN 
output DL feature 0 mean 0.015445	output DL std 1.18571
output DL feature 1 mean 0.401364	output DL std 1.26303
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      0.247      0.5391 
   1 |     -0.503       -1.16 
   2 |     0.4538      0.9789 
   3 |      -2.05      -1.808 
   4 |    -0.2042      0.2149 
   5 |   -0.07906      0.9151 
   6 |    -0.5261     -0.5578 
   7 |      1.735     0.02198 
   8 |      1.332       1.563 
   9 |    -0.4057     -0.7075 

output BN feature 0 mean -3.33067e-17	output BN std 1.05405
output BN feature 1 mean 3.33067e-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 |    0.08055    -0.09225      0.1114    0.004839 
   1 |    -0.9905      -0.165      -1.923     -0.4327 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |     0.1817       1.015      0.7095     0.04607 
   1 |     0.1833      0.2458       1.244     0.05831 

 training batch 2 mu var00.0154479
compute loss for weight  0.181757  0.181747 result 1.52383
 training batch 3 mu var00.015445
compute loss for weight  0.181737  0.181747 result 1.52383
 training batch 4 mu var00.0154457
compute loss for weight  0.181752  0.181747 result 1.52383
 training batch 5 mu var00.015445
compute loss for weight  0.181742  0.181747 result 1.52383
   --dy = 0.0805521 dy_ref = 0.0805521
 training batch 6 mu var00.0154445
compute loss for weight  1.01547  1.01546 result 1.52383
 training batch 7 mu var00.015445
compute loss for weight  1.01545  1.01546 result 1.52383
 training batch 8 mu var00.0154449
compute loss for weight  1.01546  1.01546 result 1.52383
 training batch 9 mu var00.015445
compute loss for weight  1.01545  1.01546 result 1.52383
   --dy = -0.0922472 dy_ref = -0.0922472
 training batch 10 mu var00.0154453
compute loss for weight  0.709483  0.709473 result 1.52383
 training batch 11 mu var00.015445
compute loss for weight  0.709463  0.709473 result 1.52383
 training batch 12 mu var00.0154452
compute loss for weight  0.709478  0.709473 result 1.52383
 training batch 13 mu var00.015445
compute loss for weight  0.709468  0.709473 result 1.52383
   --dy = 0.111396 dy_ref = 0.111396
 training batch 14 mu var00.015445
compute loss for weight  0.0460774  0.0460674 result 1.52383
 training batch 15 mu var00.015445
compute loss for weight  0.0460574  0.0460674 result 1.52383
 training batch 16 mu var00.015445
compute loss for weight  0.0460724  0.0460674 result 1.52383
 training batch 17 mu var00.015445
compute loss for weight  0.0460624  0.0460674 result 1.52383
   --dy = 0.00483912 dy_ref = 0.00483912
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      2.813       0.235 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.015445
compute loss for weight  1.00001  1 result 1.52385
 training batch 19 mu var00.015445
compute loss for weight  0.99999  1 result 1.5238
 training batch 20 mu var00.015445
compute loss for weight  1.00001  1 result 1.52384
 training batch 21 mu var00.015445
compute loss for weight  0.999995  1 result 1.52381
   --dy = 2.81261 dy_ref = 2.81261
 training batch 22 mu var00.015445
compute loss for weight  1.00001  1 result 1.52383
 training batch 23 mu var00.015445
compute loss for weight  0.99999  1 result 1.52382
 training batch 24 mu var00.015445
compute loss for weight  1.00001  1 result 1.52383
 training batch 25 mu var00.015445
compute loss for weight  0.999995  1 result 1.52383
   --dy = 0.235042 dy_ref = 0.235042
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -5.551e-17  -3.469e-18 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.015445
compute loss for weight  1e-05  0 result 1.52383
 training batch 27 mu var00.015445
compute loss for weight  -1e-05  0 result 1.52383
 training batch 28 mu var00.015445
compute loss for weight  5e-06  0 result 1.52383
 training batch 29 mu var00.015445
compute loss for weight  -5e-06  0 result 1.52383
   --dy = 3.70074e-11 dy_ref = -5.55112e-17
 training batch 30 mu var00.015445
compute loss for weight  1e-05  0 result 1.52383
 training batch 31 mu var00.015445
compute loss for weight  -1e-05  0 result 1.52383
 training batch 32 mu var00.015445
compute loss for weight  5e-06  0 result 1.52383
 training batch 33 mu var00.015445
compute loss for weight  -5e-06  0 result 1.52383
   --dy = 2.96059e-11 dy_ref = -3.46945e-18
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -2.463      -1.943 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     -1.142     -0.1209 

 training batch 34 mu var00.015445
compute loss for weight  -1.14171  -1.14172 result 1.5238
 training batch 35 mu var00.015445
compute loss for weight  -1.14173  -1.14172 result 1.52385
 training batch 36 mu var00.015445
compute loss for weight  -1.14171  -1.14172 result 1.52381
 training batch 37 mu var00.015445
compute loss for weight  -1.14172  -1.14172 result 1.52384
   --dy = -2.46349 dy_ref = -2.46349
 training batch 38 mu var00.015445
compute loss for weight  -0.120939  -0.120949 result 1.52381
 training batch 39 mu var00.015445
compute loss for weight  -0.120959  -0.120949 result 1.52385
 training batch 40 mu var00.015445
compute loss for weight  -0.120944  -0.120949 result 1.52382
 training batch 41 mu var00.015445
compute loss for weight  -0.120954  -0.120949 result 1.52384
   --dy = -1.94332 dy_ref = -1.94332
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m1.82967e-09[NON-XML-CHAR-0x1B][39m