Execution Time0.16s

Test: TMVA-DNN-BatchNormalization (Passed)
Build: master-x86_64-centos7-gcc62-opt-master (olhswep09.cern.ch) on 2019-11-15 11:16:19
Repository revision: ddaf537cc1431ddcb8dc5b394576c131f26a6e1a

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 var00.782987
output DL 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      1.227    -0.03781 
   1 |      2.183     -0.2972 
   2 |   -0.03608     0.09778 
   3 |     0.4356      0.3482 
   4 |     0.9208     0.08934 
   5 |      2.834    -0.02777 
   6 |     -2.785      0.6326 
   7 |      2.122     -0.9633 
   8 |      1.143     -0.1478 
   9 |    -0.2139     0.09033 

output BN 
output DL feature 0 mean 0.782987	output DL std 1.59247
output DL feature 1 mean -0.0215613	output DL std 0.419355
output of BN 

10x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |     0.2938    -0.04084 
   1 |     0.9269     -0.6926 
   2 |    -0.5422      0.2999 
   3 |    -0.2299      0.9291 
   4 |    0.09125      0.2787 
   5 |      1.357     -0.0156 
   6 |     -2.362       1.644 
   7 |     0.8864      -2.366 
   8 |     0.2381     -0.3171 
   9 |    -0.6598      0.2812 

output BN feature 0 mean -1.11022e-17	output BN std 1.05407
output BN feature 1 mean -4.44089e-17	output BN std 1.05376
Testing weight gradients   for    layer 0
weight gradient for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |   -0.04965      0.1136   -0.007325     0.02079 
   1 |       7.88       91.52       8.899       -24.8 

weights for layer 0

2x4 matrix is as follows

     |      0    |      1    |      2    |      3    |
---------------------------------------------------------
   0 |      2.064       1.079     -0.5106      -1.118 
   1 |     -0.296     -0.4764     0.07255      0.1995 

 training batch 2 mu var00.78299
compute loss for weight  2.06366  2.06365 result 6.72825
 training batch 3 mu var00.782987
compute loss for weight  2.06364  2.06365 result 6.72825
 training batch 4 mu var00.782988
compute loss for weight  2.06366  2.06365 result 6.72825
 training batch 5 mu var00.782987
compute loss for weight  2.06365  2.06365 result 6.72825
   --dy = -0.0496503 dy_ref = -0.0496503
 training batch 6 mu var00.782986
compute loss for weight  1.07863  1.07862 result 6.72825
 training batch 7 mu var00.782987
compute loss for weight  1.07861  1.07862 result 6.72825
 training batch 8 mu var00.782987
compute loss for weight  1.07863  1.07862 result 6.72825
 training batch 9 mu var00.782987
compute loss for weight  1.07862  1.07862 result 6.72825
   --dy = 0.113607 dy_ref = 0.113607
 training batch 10 mu var00.782987
compute loss for weight  -0.510638  -0.510648 result 6.72825
 training batch 11 mu var00.782987
compute loss for weight  -0.510658  -0.510648 result 6.72825
 training batch 12 mu var00.782987
compute loss for weight  -0.510643  -0.510648 result 6.72825
 training batch 13 mu var00.782987
compute loss for weight  -0.510653  -0.510648 result 6.72825
   --dy = -0.00732517 dy_ref = -0.00732517
 training batch 14 mu var00.782987
compute loss for weight  -1.11774  -1.11775 result 6.72825
 training batch 15 mu var00.782987
compute loss for weight  -1.11776  -1.11775 result 6.72825
 training batch 16 mu var00.782987
compute loss for weight  -1.11774  -1.11775 result 6.72825
 training batch 17 mu var00.782987
compute loss for weight  -1.11775  -1.11775 result 6.72825
   --dy = 0.0207904 dy_ref = 0.0207904
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      13.24      0.2153 

weights for layer 1

1x2 matrix is as follows

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

 training batch 18 mu var00.782987
compute loss for weight  1.00001  1 result 6.72838
 training batch 19 mu var00.782987
compute loss for weight  0.99999  1 result 6.72812
 training batch 20 mu var00.782987
compute loss for weight  1.00001  1 result 6.72832
 training batch 21 mu var00.782987
compute loss for weight  0.999995  1 result 6.72819
   --dy = 13.2412 dy_ref = 13.2412
 training batch 22 mu var00.782987
compute loss for weight  1.00001  1 result 6.72825
 training batch 23 mu var00.782987
compute loss for weight  0.99999  1 result 6.72825
 training batch 24 mu var00.782987
compute loss for weight  1.00001  1 result 6.72825
 training batch 25 mu var00.782987
compute loss for weight  0.999995  1 result 6.72825
   --dy = 0.215323 dy_ref = 0.215323
Testing weight gradients   for    layer 1
weight gradient for layer 1

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 | -5.551e-16   1.735e-17 

weights for layer 1

1x2 matrix is as follows

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

 training batch 26 mu var00.782987
compute loss for weight  1e-05  0 result 6.72825
 training batch 27 mu var00.782987
compute loss for weight  -1e-05  0 result 6.72825
 training batch 28 mu var00.782987
compute loss for weight  5e-06  0 result 6.72825
 training batch 29 mu var00.782987
compute loss for weight  -5e-06  0 result 6.72825
   --dy = 4.44089e-11 dy_ref = -5.55112e-16
 training batch 30 mu var00.782987
compute loss for weight  1e-05  0 result 6.72825
 training batch 31 mu var00.782987
compute loss for weight  -1e-05  0 result 6.72825
 training batch 32 mu var00.782987
compute loss for weight  5e-06  0 result 6.72825
 training batch 33 mu var00.782987
compute loss for weight  -5e-06  0 result 6.72825
   --dy = 2.51651e-10 dy_ref = 1.73472e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      5.187      -3.823 

weights for layer 2

1x2 matrix is as follows

     |      0    |      1    |
-------------------------------
   0 |      2.553    -0.05633 

 training batch 34 mu var00.782987
compute loss for weight  2.55273  2.55272 result 6.7283
 training batch 35 mu var00.782987
compute loss for weight  2.55271  2.55272 result 6.7282
 training batch 36 mu var00.782987
compute loss for weight  2.55273  2.55272 result 6.72828
 training batch 37 mu var00.782987
compute loss for weight  2.55272  2.55272 result 6.72823
   --dy = 5.18708 dy_ref = 5.18708
 training batch 38 mu var00.782987
compute loss for weight  -0.0563182  -0.0563282 result 6.72821
 training batch 39 mu var00.782987
compute loss for weight  -0.0563382  -0.0563282 result 6.72829
 training batch 40 mu var00.782987
compute loss for weight  -0.0563232  -0.0563282 result 6.72823
 training batch 41 mu var00.782987
compute loss for weight  -0.0563332  -0.0563282 result 6.72827
   --dy = -3.82266 dy_ref = -3.82266
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][33m2.84746e-08[NON-XML-CHAR-0x1B][39m