Train accuracy = 89.648%

per_face_accuracy: 0.8964884874006657
mean_confidence: 0.90162826
class_0_acc: 0.8309014514896867
class_1_acc: 0.9787369493251846
class_2_acc: 0.9244204509368054
class_3_acc: 0.9623172103487064
class_4_acc: 0.9210676971518534
class_5_acc: 0.8517195869885879
class_6_acc: 0.9219792629093522
per_class_accuracy: 0.9130203727357395
IoU: 0.80784581826577
1515/1515 [00:53<00:00, 28.31it/s]



Test Accuracy = 86.432%

per_face_accuracy: 0.864323620706919
mean_confidence: 0.89635736
class_0_acc: 0.7909681833732467
class_1_acc: 0.9679347826086957
class_2_acc: 0.9298013245033112
class_3_acc: 0.8977035490605428
class_4_acc: 0.9051359516616314
class_5_acc: 0.7790178571428571
class_6_acc: 0.8698284561049445
per_class_accuracy: 0.8771985863507469
IoU: 0.7578926103706184
191/191 [00:10<00:00, 18.96it/s]



Unseen Accuracy (Current AI Model) = 86.64%

per_face_accuracy: 0.866403162055336
mean_confidence: 0.8827821
class_0_acc: 0.875
class_1_acc: 0.9848156182212582
class_2_acc: 0.839622641509434
class_3_acc: 0.9245283018867925
class_4_acc: 0.7870722433460076
class_5_acc: 0.8002853067047075
class_6_acc: 0.9120370370370371
per_class_accuracy: 0.8747658783864625
IoU: 0.7264917818820532
49/49 [00:03<00:00, 13.94it/s]


Unseen Accuracy (Previous AI Model) = 64.15%

per_face_accuracy: 0.6415019762845849
mean_confidence: 0.79006696
class_0_acc: 0.5881410256410257
class_1_acc: 0.93058568329718
class_2_acc: 0.38207547169811323
class_3_acc: 0.22641509433962265
class_4_acc: 0.6273764258555133
class_5_acc: 0.6433666191155493
class_6_acc: 0.5462962962962963
per_class_accuracy: 0.5634652308919
IoU: 0.40713622181722414
49/49 [00:08<00:00,  6.12it/s]