Execution Time0.11s

Test: TMVA-DNN-BatchNormalization-Cpu (Passed)
Build: master-aarch64-centos7-gcc48 (techlab-arm64-moonshot-xgene-004) on 2019-11-14 00:49:23
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 var00.856468
output DL 
output BN 
output DL feature 0 mean 0.856468	output DL std 1.71855
output DL feature 1 mean -0.749037	output DL std 1.42762
output of BN 
output BN feature 0 mean -6.66134e-17	output BN std 1.05407
output BN feature 1 mean 2.22045e-17	output BN std 1.05406
Testing weight gradients   for    layer 0
weight gradient for layer 0
weights for layer 0
 training batch 2 mu var00.856471
compute loss for weight  1.70115  1.70114 result 0.859817
 training batch 3 mu var00.856468
compute loss for weight  1.70113  1.70114 result 0.859818
 training batch 4 mu var00.856469
compute loss for weight  1.70114  1.70114 result 0.859817
 training batch 5 mu var00.856468
compute loss for weight  1.70113  1.70114 result 0.859817
   --dy = -0.0585339 dy_ref = -0.0585339
 training batch 6 mu var00.856467
compute loss for weight  0.948072  0.948062 result 0.859818
 training batch 7 mu var00.856468
compute loss for weight  0.948052  0.948062 result 0.859816
 training batch 8 mu var00.856468
compute loss for weight  0.948067  0.948062 result 0.859818
 training batch 9 mu var00.856468
compute loss for weight  0.948057  0.948062 result 0.859817
   --dy = 0.119953 dy_ref = 0.119953
 training batch 10 mu var00.856468
compute loss for weight  0.573195  0.573185 result 0.859817
 training batch 11 mu var00.856468
compute loss for weight  0.573175  0.573185 result 0.859817
 training batch 12 mu var00.856468
compute loss for weight  0.57319  0.573185 result 0.859817
 training batch 13 mu var00.856468
compute loss for weight  0.57318  0.573185 result 0.859817
   --dy = -0.0295732 dy_ref = -0.0295732
 training batch 14 mu var00.856468
compute loss for weight  -0.0275828  -0.0275928 result 0.859816
 training batch 15 mu var00.856468
compute loss for weight  -0.0276028  -0.0275928 result 0.859818
 training batch 16 mu var00.856468
compute loss for weight  -0.0275878  -0.0275928 result 0.859817
 training batch 17 mu var00.856468
compute loss for weight  -0.0275978  -0.0275928 result 0.859818
   --dy = -0.101196 dy_ref = -0.101196
Testing weight gradients   for    layer 1
weight gradient for layer 1
weights for layer 1
 training batch 18 mu var00.856468
compute loss for weight  1.00001  1 result 0.859815
 training batch 19 mu var00.856468
compute loss for weight  0.99999  1 result 0.85982
 training batch 20 mu var00.856468
compute loss for weight  1.00001  1 result 0.859816
 training batch 21 mu var00.856468
compute loss for weight  0.999995  1 result 0.859818
   --dy = -0.263536 dy_ref = -0.263536
 training batch 22 mu var00.856468
compute loss for weight  1.00001  1 result 0.859837
 training batch 23 mu var00.856468
compute loss for weight  0.99999  1 result 0.859797
 training batch 24 mu var00.856468
compute loss for weight  1.00001  1 result 0.859827
 training batch 25 mu var00.856468
compute loss for weight  0.999995  1 result 0.859807
   --dy = 1.98317 dy_ref = 1.98317
Testing weight gradients   for    layer 1
weight gradient for layer 1
weights for layer 1
 training batch 26 mu var00.856468
compute loss for weight  1e-05  0 result 0.859817
 training batch 27 mu var00.856468
compute loss for weight  -1e-05  0 result 0.859817
 training batch 28 mu var00.856468
compute loss for weight  5e-06  0 result 0.859817
 training batch 29 mu var00.856468
compute loss for weight  -5e-06  0 result 0.859817
   --dy = 0 dy_ref = 0
 training batch 30 mu var00.856468
compute loss for weight  1e-05  0 result 0.859817
 training batch 31 mu var00.856468
compute loss for weight  -1e-05  0 result 0.859817
 training batch 32 mu var00.856468
compute loss for weight  5e-06  0 result 0.859817
 training batch 33 mu var00.856468
compute loss for weight  -5e-06  0 result 0.859817
   --dy = 0 dy_ref = -2.77556e-17
Testing weight gradients   for    layer 2
weight gradient for layer 2
weights for layer 2
 training batch 34 mu var00.856468
compute loss for weight  0.22091  0.2209 result 0.859805
 training batch 35 mu var00.856468
compute loss for weight  0.22089  0.2209 result 0.859829
 training batch 36 mu var00.856468
compute loss for weight  0.220905  0.2209 result 0.859811
 training batch 37 mu var00.856468
compute loss for weight  0.220895  0.2209 result 0.859823
   --dy = -1.19301 dy_ref = -1.19301
 training batch 38 mu var00.856468
compute loss for weight  1.0827  1.08269 result 0.859835
 training batch 39 mu var00.856468
compute loss for weight  1.08268  1.08269 result 0.859799
 training batch 40 mu var00.856468
compute loss for weight  1.08269  1.08269 result 0.859826
 training batch 41 mu var00.856468
compute loss for weight  1.08268  1.08269 result 0.859808
   --dy = 1.83171 dy_ref = 1.83171
Testing weight gradients:      maximum relative error: [NON-XML-CHAR-0x1B][32m6.10509e-10[NON-XML-CHAR-0x1B][39m