2019-10-25 17:36:27.481900: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cublas64_100.dll
Train on 48000 samples, validate on 12000 samples
Epoch 1/30
48000/48000 [==============================] - 3s 59us/sample - loss: 0.4291 - accuracy: 0.8689 - val_loss: 0.1768 - val_accuracy: 0.9467
Epoch 2/30
48000/48000 [==============================] - 2s 34us/sample - loss: 0.1978 - accuracy: 0.9411 - val_loss: 0.1277 - val_accuracy: 0.9629
Epoch 3/30
48000/48000 [==============================] - 2s 31us/sample - loss: 0.1560 - accuracy: 0.9538 - val_loss: 0.1061 - val_accuracy: 0.9701
Epoch 4/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.1298 - accuracy: 0.9616 - val_loss: 0.0998 - val_accuracy: 0.9735
Epoch 5/30
48000/48000 [==============================] - 2s 31us/sample - loss: 0.1154 - accuracy: 0.9664 - val_loss: 0.0977 - val_accuracy: 0.9719
Epoch 6/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.1042 - accuracy: 0.9697 - val_loss: 0.0912 - val_accuracy: 0.9771
Epoch 7/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0973 - accuracy: 0.9714 - val_loss: 0.0918 - val_accuracy: 0.9778
Epoch 8/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0902 - accuracy: 0.9735 - val_loss: 0.0953 - val_accuracy: 0.9758
Epoch 9/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0833 - accuracy: 0.9757 - val_loss: 0.0913 - val_accuracy: 0.9772
Epoch 10/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0810 - accuracy: 0.9762 - val_loss: 0.0923 - val_accuracy: 0.9791
Epoch 11/30
48000/48000 [==============================] - 2s 32us/sample - loss: 0.0758 - accuracy: 0.9778 - val_loss: 0.0947 - val_accuracy: 0.9779
Epoch 12/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0762 - accuracy: 0.9777 - val_loss: 0.1020 - val_accuracy: 0.9762
Epoch 13/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0724 - accuracy: 0.9787 - val_loss: 0.0987 - val_accuracy: 0.9781
Epoch 14/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0695 - accuracy: 0.9804 - val_loss: 0.0949 - val_accuracy: 0.9793
Epoch 15/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0650 - accuracy: 0.9810 - val_loss: 0.0964 - val_accuracy: 0.9811
Epoch 16/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0657 - accuracy: 0.9815 - val_loss: 0.1021 - val_accuracy: 0.9793
Epoch 17/30
48000/48000 [==============================] - 2s 31us/sample - loss: 0.0627 - accuracy: 0.9825 - val_loss: 0.1033 - val_accuracy: 0.9804
Epoch 18/30
48000/48000 [==============================] - 2s 31us/sample - loss: 0.0622 - accuracy: 0.9822 - val_loss: 0.1056 - val_accuracy: 0.9798
Epoch 19/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0608 - accuracy: 0.9828 - val_loss: 0.1027 - val_accuracy: 0.9796
Epoch 20/30
48000/48000 [==============================] - 2s 31us/sample - loss: 0.0622 - accuracy: 0.9829 - val_loss: 0.1042 - val_accuracy: 0.9799
Epoch 21/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0573 - accuracy: 0.9845 - val_loss: 0.1108 - val_accuracy: 0.9797
Epoch 22/30
48000/48000 [==============================] - 2s 32us/sample - loss: 0.0568 - accuracy: 0.9846 - val_loss: 0.1111 - val_accuracy: 0.9797
Epoch 23/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0559 - accuracy: 0.9848 - val_loss: 0.1088 - val_accuracy: 0.9797
Epoch 24/30
48000/48000 [==============================] - 2s 32us/sample - loss: 0.0535 - accuracy: 0.9847 - val_loss: 0.1214 - val_accuracy: 0.9774
Epoch 25/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0557 - accuracy: 0.9844 - val_loss: 0.1185 - val_accuracy: 0.9801
Epoch 26/30
48000/48000 [==============================] - 2s 31us/sample - loss: 0.0487 - accuracy: 0.9864 - val_loss: 0.1240 - val_accuracy: 0.9806
Epoch 27/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0559 - accuracy: 0.9849 - val_loss: 0.1203 - val_accuracy: 0.9803
Epoch 28/30
48000/48000 [==============================] - 2s 31us/sample - loss: 0.0516 - accuracy: 0.9861 - val_loss: 0.1198 - val_accuracy: 0.9797
Epoch 29/30
48000/48000 [==============================] - 1s 31us/sample - loss: 0.0512 - accuracy: 0.9863 - val_loss: 0.1145 - val_accuracy: 0.9817
Epoch 30/30
48000/48000 [==============================] - 2s 31us/sample - loss: 0.0525 - accuracy: 0.9856 - val_loss: 0.1233 - val_accuracy: 0.9801