Hi, I build the Dence model with Keras. After tuning model, but the val_acc of best model is not the same as the best val_acc of epoch. one is 0.7945351312987935, the other is 0.7932487579843861. Why it happen this problem?
Best validation acc of epoch:
0.7945351312987935
100%|██████████| 3/3 [00:46<00:00, 16.57s/it, best loss: -0.7945351312987935]
Best performing model chosen hyper-parameters:
{'batch_size': 2, 'units': 1, 'units_1': 1, 'units_2': 0, 'units_3': 0, 'units_4': 0, 'units_5': 1}
Evalutation of best performing model:
32/22544 [..............................] - ETA: 0s
4640/22544 [=====>........................] - ETA: 0s
9632/22544 [===========>..................] - ETA: 0s
14592/22544 [==================>...........] - ETA: 0s
19456/22544 [========================>.....] - ETA: 0s
22544/22544 [==============================] - 0s 11us/step
[1.4304323981551095, 0.7932487579843861]
Hi, I build the Dence model with Keras. After tuning model, but the val_acc of best model is not the same as the best val_acc of epoch. one is
0.7945351312987935
, the other is0.7932487579843861
. Why it happen this problem?