Train on 1291 samples, validate on 144 samples
Epoch 1/20
2019-01-19 23:28:59.243532: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1511] Adding visible gpu devices: 0
2019-01-19 23:28:59.243565: I tensorflow/core/common_runtime/gpu/gpu_device.cc:982] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-01-19 23:28:59.243571: I tensorflow/core/common_runtime/gpu/gpu_device.cc:988]      0
2019-01-19 23:28:59.243575: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1001] 0:   N
2019-01-19 23:28:59.243744: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 7823 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:01:00.0, compute capability: 6.1)
1291/1291 [==============================] - 2s 2ms/step - loss: 0.3498 - acc: 0.8397 - val_loss: 0.2467 - val_acc: 0.8778
Epoch 2/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.2105 - acc: 0.8984 - val_loss: 0.1924 - val_acc: 0.9069
Epoch 3/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.1663 - acc: 0.9247 - val_loss: 0.1694 - val_acc: 0.9194
Epoch 4/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.1389 - acc: 0.9421 - val_loss: 0.1562 - val_acc: 0.9333
Epoch 5/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.1175 - acc: 0.9554 - val_loss: 0.1399 - val_acc: 0.9486
Epoch 6/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.1025 - acc: 0.9603 - val_loss: 0.1276 - val_acc: 0.9514
Epoch 7/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0916 - acc: 0.9655 - val_loss: 0.1156 - val_acc: 0.9639
Epoch 8/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0833 - acc: 0.9695 - val_loss: 0.1069 - val_acc: 0.9681
Epoch 9/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0762 - acc: 0.9730 - val_loss: 0.1023 - val_acc: 0.9750
Epoch 10/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0692 - acc: 0.9747 - val_loss: 0.0952 - val_acc: 0.9750
Epoch 11/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0634 - acc: 0.9768 - val_loss: 0.0892 - val_acc: 0.9750
Epoch 12/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0580 - acc: 0.9785 - val_loss: 0.0788 - val_acc: 0.9792
Epoch 13/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0533 - acc: 0.9786 - val_loss: 0.0718 - val_acc: 0.9806
Epoch 14/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0489 - acc: 0.9811 - val_loss: 0.0665 - val_acc: 0.9806
Epoch 15/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0450 - acc: 0.9830 - val_loss: 0.0631 - val_acc: 0.9806
Epoch 16/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0421 - acc: 0.9842 - val_loss: 0.0606 - val_acc: 0.9806
Epoch 17/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0386 - acc: 0.9862 - val_loss: 0.0594 - val_acc: 0.9778
Epoch 18/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0363 - acc: 0.9867 - val_loss: 0.0567 - val_acc: 0.9792
Epoch 19/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0337 - acc: 0.9876 - val_loss: 0.0556 - val_acc: 0.9806
Epoch 20/20
1291/1291 [==============================] - 1s 1ms/step - loss: 0.0320 - acc: 0.9884 - val_loss: 0.0534 - val_acc: 0.9819
1291/1291 [==============================] - 0s 67us/step
Test sample 44, actual label is [1 0 0 0 1] but predicted as [1 0 0 0 0]
Test sample 52, actual label is [1 0 0 0 0] but predicted as [1 0 0 1 0]
Test sample 55, actual label is [1 0 0 0 1] but predicted as [1 0 0 0 0]
Test sample 57, actual label is [1 0 0 0 1] but predicted as [1 0 0 0 0]
Test sample 69, actual label is [0 0 0 0 1] but predicted as [1 0 0 0 1]
Test sample 78, actual label is [0 0 0 0 1] but predicted as [1 0 0 0 1]
Test sample 87, actual label is [0 0 0 1 0] but predicted as [0 0 0 0 0]
Test sample 98, actual label is [1 0 0 0 1] but predicted as [1 0 0 1 0]
Test sample 120, actual label is [1 0 0 1 0] but predicted as [1 0 0 0 0]
Test sample 128, actual label is [1 0 0 0 1] but predicted as [1 0 0 1 0]
Test sample 133, actual label is [1 0 0 0 0] but predicted as [1 0 0 1 0]
Accuracy is 0.9236
testing done