EIP_2
Logs for 20 epochs
Accuracy >= 99.40 Achieved in EPOC 8, 16, 18, 19, 20
Max Accuracy = 99.46
Epoch 00001: LearningRateScheduler setting learning rate to 0.003. 60000/60000 [==============================] - 16s 274us/step - loss: 0.5366 - acc: 0.8495 - val_loss: 0.1032 - val_acc: 0.9799 Epoch 2/20
Epoch 00002: LearningRateScheduler setting learning rate to 0.0022744503. 60000/60000 [==============================] - 11s 180us/step - loss: 0.2529 - acc: 0.9262 - val_loss: 0.0600 - val_acc: 0.9877 Epoch 3/20
Epoch 00003: LearningRateScheduler setting learning rate to 0.0018315018. 60000/60000 [==============================] - 11s 185us/step - loss: 0.2008 - acc: 0.9394 - val_loss: 0.0494 - val_acc: 0.9873 Epoch 4/20
Epoch 00004: LearningRateScheduler setting learning rate to 0.0015329586. 60000/60000 [==============================] - 11s 179us/step - loss: 0.1719 - acc: 0.9452 - val_loss: 0.0433 - val_acc: 0.9888 Epoch 5/20
Epoch 00005: LearningRateScheduler setting learning rate to 0.0013181019. 60000/60000 [==============================] - 11s 182us/step - loss: 0.1516 - acc: 0.9514 - val_loss: 0.0443 - val_acc: 0.9864 Epoch 6/20
Epoch 00006: LearningRateScheduler setting learning rate to 0.0011560694. 60000/60000 [==============================] - 11s 178us/step - loss: 0.1439 - acc: 0.9501 - val_loss: 0.0309 - val_acc: 0.9913 Epoch 7/20
Epoch 00007: LearningRateScheduler setting learning rate to 0.0010295127. 60000/60000 [==============================] - 11s 183us/step - loss: 0.1302 - acc: 0.9528 - val_loss: 0.0251 - val_acc: 0.9922 Epoch 8/20
Epoch 00008: LearningRateScheduler setting learning rate to 0.0009279307. 60000/60000 [==============================] - 11s 180us/step - loss: 0.1233 - acc: 0.9539 - val_loss: 0.0230 - val_acc: 0.9940 Epoch 9/20
Epoch 00009: LearningRateScheduler setting learning rate to 0.0008445946. 60000/60000 [==============================] - 11s 182us/step - loss: 0.1184 - acc: 0.9553 - val_loss: 0.0254 - val_acc: 0.9923 Epoch 10/20
Epoch 00010: LearningRateScheduler setting learning rate to 0.0007749935. 60000/60000 [==============================] - 11s 179us/step - loss: 0.1162 - acc: 0.9533 - val_loss: 0.0261 - val_acc: 0.9917 Epoch 11/20
Epoch 00011: LearningRateScheduler setting learning rate to 0.0007159905. 60000/60000 [==============================] - 11s 180us/step - loss: 0.1066 - acc: 0.9575 - val_loss: 0.0232 - val_acc: 0.9932 Epoch 12/20
Epoch 00012: LearningRateScheduler setting learning rate to 0.000665336. 60000/60000 [==============================] - 11s 182us/step - loss: 0.1064 - acc: 0.9554 - val_loss: 0.0220 - val_acc: 0.9930 Epoch 13/20
Epoch 00013: LearningRateScheduler setting learning rate to 0.0006213753. 60000/60000 [==============================] - 11s 179us/step - loss: 0.1027 - acc: 0.9576 - val_loss: 0.0216 - val_acc: 0.9934 Epoch 14/20
Epoch 00014: LearningRateScheduler setting learning rate to 0.0005828638. 60000/60000 [==============================] - 11s 181us/step - loss: 0.1021 - acc: 0.9561 - val_loss: 0.0210 - val_acc: 0.9927 Epoch 15/20
Epoch 00015: LearningRateScheduler setting learning rate to 0.0005488474. 60000/60000 [==============================] - 11s 181us/step - loss: 0.1010 - acc: 0.9564 - val_loss: 0.0197 - val_acc: 0.9938 Epoch 16/20
Epoch 00016: LearningRateScheduler setting learning rate to 0.0005185825. 60000/60000 [==============================] - 11s 184us/step - loss: 0.0962 - acc: 0.9583 - val_loss: 0.0194 - val_acc: 0.9943 Epoch 17/20
Epoch 00017: LearningRateScheduler setting learning rate to 0.000491481. 60000/60000 [==============================] - 11s 181us/step - loss: 0.0973 - acc: 0.9566 - val_loss: 0.0202 - val_acc: 0.9932 Epoch 18/20
Epoch 00018: LearningRateScheduler setting learning rate to 0.0004670715. 60000/60000 [==============================] - 11s 181us/step - loss: 0.0947 - acc: 0.9578 - val_loss: 0.0182 - val_acc: 0.9946 Epoch 19/20
Epoch 00019: LearningRateScheduler setting learning rate to 0.0004449718. 60000/60000 [==============================] - 11s 179us/step - loss: 0.0937 - acc: 0.9581 - val_loss: 0.0183 - val_acc: 0.9944 Epoch 20/20
Epoch 00020: LearningRateScheduler setting learning rate to 0.000424869. 60000/60000 [==============================] - 11s 177us/step - loss: 0.0909 - acc: 0.9578 - val_loss: 0.0191 - val_acc: 0.9940 <keras.callbacks.History at 0x7fc862761da0>
Model.evalutate Result :
[0.01914663181611104, 0.994]
Strategy Used:
Reduced number of parameters by reducing the kernel size.