Create and test proper validation set about cancerdetection

[JoshVarty]

All transforms on Resnet50

Epochs	Valid	Test
10	0.935	0.9612
20	0.938	0.9624
30	0.934	0.9628
40	0.940	0.9640
50	0.938	0.9637
60	0.941	0.9654
70	0.937	0.9645
100*	0.939	0.9686
160*	0.945	0.9690

* Trained using .fit()

[JoshVarty]

All transforms on Resnet101

Epochs	Valid	Test
10	0.884	0.9412
20	0.898	0.9440
30	0.901	0.9469
40	0.901	0.9492
50	0.904	0.9496
60	0.908	0.9522
70	0.916	0.9530
100*	0.924	0.9555
160*	0.937	0.9599

[JoshVarty]

All transforms on Resnet50 (random validation set)

Epochs	Valid	Test
10	0.986	0.9612
20	0.987	0.9604
30	0.989	0.9603
40	0.990	0.9601
50	0.990	0.9600
60	0.991	0.9611
70	0.992	0.9622
100*	0.992	0.9592
160*	0.994	0.9578

* Trained using .fit()