iCaRL - one epoch

Mixing the classes and putting them in batches of classes...
Creating a validation set ...
Batch of classes number 1 arrives ...
Batch of classes 1 out of 10 batches
Epoch 0
0.287272
0.00871001
0.00892173
0.00863384
0.00751852
0.00657171
0.00608974
0.00585746
Training accuracy 0.062500
0.0057788
0.00560296
0.00555174
0.00548918
0.0055367
0.00539182
0.00533972
0.00537566
Training accuracy 0.070312
0.00534246
0.00530634
0.00524906
0.00524764
0.00522196
0.00521024
0.00523701
0.00517737
Training accuracy 0.078125
0.005154
0.00515257
0.00507729
0.0051164
0.0050586
0.00501388
0.00512017
0.00504586
Training accuracy 0.085938
0.00494739
0.00499304
0.00497555
0.00495446
0.00496844
0.00494343
0.00493757
0.00485029
Training accuracy 0.140625
0.00482917
0.00483164
0.00488351
0.00479883
0.00483863
0.00484456
0.0048152
0.00476418
Training accuracy 0.078125
0.00475501
0.00481238
0.00475885
0.00478915
0.00470401
0.00471081
0.00468985
0.00466995
Training accuracy 0.132812
0.00463978
0.00475154
0.00460605
0.0046251
0.00460757
0.00460384
0.00455722
0.00461336
Training accuracy 0.156250
0.00457644
0.00455298
0.00454285
0.0045895
0.00455414
0.00451121
0.00456292
0.0045069
Training accuracy 0.187500
0.00452433
0.00448638
0.00447541
0.00442919
0.00447134
0.0044896
0.0044824
0.00450531
Training accuracy 0.195312
0.00448563
0.00442603
0.00439628
0.00439986
0.00439687
0.00440562
0.00435975
0.00443868
Training accuracy 0.187500
0.00435689
0.00437406
0.00437823
0.00437819
0.00429873
0.00438045
0.00433074
0.0043012
Training accuracy 0.242188
0.00433212
Exemplars selection starting ...
Computing theoretical class means for NCM and mean-of-exemplars for iCaRL ...
Batch of classes number 2 arrives ...
Batch of classes 2 out of 10 batches
Epoch 0
0.011735
0.0121078
0.0111618
0.010403
0.00974831
0.00949552
0.00924931
0.00905704
Training accuracy 0.062500
0.00905689
0.00911064
0.00898644
0.00896149
0.00897088
0.00886607
0.00886964
0.00882033
Training accuracy 0.046875
0.0088617
0.00889832
0.00885294
0.0086567
0.00872421
0.0087801
0.00876649
0.00870548
Training accuracy 0.085938
0.00875134
0.00867434
0.00867335
0.00865952
0.0087811
0.00874177
0.00869078
0.00861675
Training accuracy 0.101562
0.00877293
0.00871357
0.00860829
0.00870467
0.00875316
0.00861962
0.00868724
0.00861359
Training accuracy 0.078125
0.00863145
0.00872596
0.0085802
0.00867081
0.00861635
0.00861645
0.00867638
0.00864936
Training accuracy 0.132812
0.00862112
0.00855318
0.00865119
0.00867549
0.00866487
0.00859771
0.00850684
0.00858965
Training accuracy 0.140625
0.00859381
0.00856225
0.00861783
0.00853969
0.00860939
0.00851869
0.00861566
0.00856602
Training accuracy 0.078125
0.00856246
0.00847558
0.0086258
0.00857576
0.00860878
0.00861267
0.00855802
0.00849134
Training accuracy 0.148438
0.00858278
0.00856503
0.00848093
0.00856737
0.00860464
0.0084817
0.0085678
0.00853012
Training accuracy 0.078125
0.00852588
0.00850274
0.00844964
0.00855854
0.00853108
0.00853306
0.00862732
0.00848812
Training accuracy 0.179688
0.00852059
0.00847373
0.00846274
0.00853646
0.00850241
0.00846972
0.00858296
0.00844549
Training accuracy 0.203125
0.00850263
0.00846575
0.00849627
0.00847889
0.0084808
0.00847772
0.00846598
0.00856872
Training accuracy 0.156250
0.00856192
0.00842343
0.00854151
0.00842551
0.00848143
0.00842469
0.00835394
Exemplars selection starting ...
Computing theoretical class means for NCM and mean-of-exemplars for iCaRL ...
Batch of classes number 3 arrives ...
Batch of classes 3 out of 10 batches
Epoch 0
0.0133068
0.0160411
0.0152802
0.0145554
0.0138945
0.0134506
0.013309
0.0132363
Training accuracy 0.054688
0.0131409
0.0129834
0.0129979
0.0130395
0.0129674
0.012992
0.0129019
0.012907
Training accuracy 0.046875
0.0129227
0.0129192
0.0128864
0.0128227
0.0128627
0.0127671
0.0127762
0.012841
Training accuracy 0.085938
0.0127621
0.0127453
0.0127394
0.0127804
0.0127349
0.0127909
0.0126875
0.012652
Training accuracy 0.125000
0.0127161
0.0127041
0.0126931
0.0127085
0.0126899
0.0126107
0.0127092
0.0126483
Training accuracy 0.093750
0.0127544
0.0126553
0.0127003
0.0125799
0.0126261
0.0125648
0.0126154
0.0126282
Training accuracy 0.132812
0.0126254
0.0127322
0.0126617
0.0126598
0.0125445
0.0126811
0.0125878
0.0126139
Training accuracy 0.132812
0.0126788
0.0126547
0.0126342
0.0125765
0.0125175
0.0125667
0.0124526
0.0125487
Training accuracy 0.195312
0.0125424
0.0126633
0.0125563
0.0125391
0.0126417
0.0125652
0.0125713
0.0125582
Training accuracy 0.140625
0.0125737
0.012591
0.0125367
0.0125514
0.0125518
0.012547
0.0125682
0.0125115
Training accuracy 0.179688
0.0125117
0.0126273
0.0126121
0.0125133
0.0125625
0.0125111
0.0124418
0.0125072
Training accuracy 0.140625
0.0124712
0.0125143
0.0125299
0.0124889
0.0125016
0.0126104
0.0125008
0.0125031
Training accuracy 0.140625
0.0124458
0.0124321
0.0125703
0.0124474
0.0125129
0.0124093
0.0125206
0.0124371
Training accuracy 0.156250
0.0123995
0.012514
0.0124937
0.0124761
0.0124713
0.0124752
0.0125337
0.0125326
Training accuracy 0.226562
Exemplars selection starting ...
Computing theoretical class means for NCM and mean-of-exemplars for iCaRL ...
Batch of classes number 4 arrives ...
Batch of classes 4 out of 10 batches
Epoch 0
0.0150059
0.0206414
0.0200337
0.0192729
0.0186985
0.0182909
0.017981
0.0178339
Training accuracy 0.015625
0.0176549
0.017499
0.0173934
0.0175482
0.017556
0.01739
0.0173688
0.0174655
Training accuracy 0.015625
0.0174182
0.0174006
0.0173555
0.0173084
0.0172947
0.0173323
0.0173499
0.0172396
Training accuracy 0.015625
0.017452
0.0173198
0.0172841
0.0172596
0.0172748
0.0173
0.0172711
0.0173558
Training accuracy 0.062500
0.0172768
0.0172309
0.0172989
0.0173261
0.0172818
0.0172231
0.0172613
0.0173507
Training accuracy 0.117188
0.017095
0.0172337
0.0172449
0.0171851
0.0171699
0.0172261
0.0171296
0.0171692
Training accuracy 0.132812
0.0172217
0.0172001
0.0172492
0.0171694
0.0170979
0.0172475
0.01708
0.0172318
Training accuracy 0.062500
0.0171684
0.0170606
0.0171323
0.017094
0.0170702
0.0171387
0.0171015
0.0172059
Training accuracy 0.117188
0.0170693
0.0171517
0.0172013
0.0172022
0.0172451
0.0171009