Execution Time0.17s

Test: TMVA-DNN-BatchNormalization-Cpu (Passed)
Build: master-x86_64-centos7-gcc62-opt-no-rt-cxxmodules (olsnba08.cern.ch) on 2019-11-14 01:02:24
Repository revision: 32b17abcda23e44b64218a42d0ca69cb30cda7e0

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 
 training batch 1 mu var00.0944043
output DL 
output BN 
output DL feature 0 mean 0.0944043	output DL std 0.707226
output DL feature 1 mean -0.111869	output DL std 0.577142
output of BN 
output BN feature 0 mean -4.44089e-17	output BN std 1.05398
output BN feature 1 mean 1.11022e-17	output BN std 1.05392
Testing weight gradients   for    layer 0
weight gradient for layer 0
weights for layer 0
 training batch 2 mu var00.0944072
compute loss for weight  0.210869  0.210859 result 0.504115
 training batch 3 mu var00.0944043
compute loss for weight  0.210849  0.210859 result 0.504116
 training batch 4 mu var00.0944051
compute loss for weight  0.210864  0.210859 result 0.504115
 training batch 5 mu var00.0944043
compute loss for weight  0.210854  0.210859 result 0.504116
   --dy = -0.0496339 dy_ref = -0.0496339
 training batch 6 mu var00.0944039
compute loss for weight  0.383615  0.383605 result 0.504115
 training batch 7 mu var00.0944043
compute loss for weight  0.383595  0.383605 result 0.504116
 training batch 8 mu var00.0944042
compute loss for weight  0.38361  0.383605 result 0.504115
 training batch 9 mu var00.0944043
compute loss for weight  0.3836  0.383605 result 0.504116
   --dy = -0.0348362 dy_ref = -0.0348362
 training batch 10 mu var00.0944046
compute loss for weight  0.379696  0.379686 result 0.504117
 training batch 11 mu var00.0944043
compute loss for weight  0.379676  0.379686 result 0.504114
 training batch 12 mu var00.0944045
compute loss for weight  0.379691  0.379686 result 0.504116
 training batch 13 mu var00.0944043
compute loss for weight  0.379681  0.379686 result 0.504115
   --dy = 0.122966 dy_ref = 0.122966
 training batch 14 mu var00.0944043
compute loss for weight  0.334832  0.334822 result 0.504115
 training batch 15 mu var00.0944043
compute loss for weight  0.334812  0.334822 result 0.504116
 training batch 16 mu var00.0944043
compute loss for weight  0.334827  0.334822 result 0.504115
 training batch 17 mu var00.0944043
compute loss for weight  0.334817  0.334822 result 0.504116
   --dy = -0.0674462 dy_ref = -0.0674462
Testing weight gradients   for    layer 1
weight gradient for layer 1
weights for layer 1
 training batch 18 mu var00.0944043
compute loss for weight  1.00001  1 result 0.504128
 training batch 19 mu var00.0944043
compute loss for weight  0.99999  1 result 0.504103
 training batch 20 mu var00.0944043
compute loss for weight  1.00001  1 result 0.504122
 training batch 21 mu var00.0944043
compute loss for weight  0.999995  1 result 0.504109
   --dy = 1.24773 dy_ref = 1.24773
 training batch 22 mu var00.0944043
compute loss for weight  1.00001  1 result 0.504113
 training batch 23 mu var00.0944043
compute loss for weight  0.99999  1 result 0.504118
 training batch 24 mu var00.0944043
compute loss for weight  1.00001  1 result 0.504114
 training batch 25 mu var00.0944043
compute loss for weight  0.999995  1 result 0.504117
   --dy = -0.239495 dy_ref = -0.239495
Testing weight gradients   for    layer 1
weight gradient for layer 1
weights for layer 1
 training batch 26 mu var00.0944043
compute loss for weight  1e-05  0 result 0.504115
 training batch 27 mu var00.0944043
compute loss for weight  -1e-05  0 result 0.504115
 training batch 28 mu var00.0944043
compute loss for weight  5e-06  0 result 0.504115
 training batch 29 mu var00.0944043
compute loss for weight  -5e-06  0 result 0.504115
   --dy = 2.96059e-11 dy_ref = -1.66533e-16
 training batch 30 mu var00.0944043
compute loss for weight  1e-05  0 result 0.504115
 training batch 31 mu var00.0944043
compute loss for weight  -1e-05  0 result 0.504115
 training batch 32 mu var00.0944043
compute loss for weight  5e-06  0 result 0.504115
 training batch 33 mu var00.0944043
compute loss for weight  -5e-06  0 result 0.504115
   --dy = -1.66533e-11 dy_ref = -2.77556e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2
weights for layer 2
 training batch 34 mu var00.0944043
compute loss for weight  0.886907  0.886897 result 0.50413
 training batch 35 mu var00.0944043
compute loss for weight  0.886887  0.886897 result 0.504101
 training batch 36 mu var00.0944043
compute loss for weight  0.886902  0.886897 result 0.504122
 training batch 37 mu var00.0944043
compute loss for weight  0.886892  0.886897 result 0.504108
   --dy = 1.40684 dy_ref = 1.40684
 training batch 38 mu var00.0944043
compute loss for weight  0.20693  0.20692 result 0.504104
 training batch 39 mu var00.0944043
compute loss for weight  0.20691  0.20692 result 0.504127
 training batch 40 mu var00.0944043
compute loss for weight  0.206925  0.20692 result 0.50411
 training batch 41 mu var00.0944043
compute loss for weight  0.206915  0.20692 result 0.504121
   --dy = -1.15743 dy_ref = -1.15743
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m5.21997e-10[NON-XML-CHAR-0x1B][39m