python cnn 实例_CNN 可视化示例

Train on 60000 samples, validate on 10000 samples

Epoch 0

60000/60000 [==============================] - 79s - loss: 0.2596 - acc: 0.9200 - val_loss: 0.0548 - val_acc: 0.9826

Epoch 1

60000/60000 [==============================] - 79s - loss: 0.0961 - acc: 0.9713 - val_loss: 0.0441 - val_acc: 0.9861

Epoch 2

60000/60000 [==============================] - 79s - loss: 0.0735 - acc: 0.9782 - val_loss: 0.0426 - val_acc: 0.9860

Epoch 3

60000/60000 [==============================] - 79s - loss: 0.0617 - acc: 0.9816 - val_loss: 0.0330 - val_acc: 0.9885

Epoch 4

60000/60000 [==============================] - 79s - loss: 0.0513 - acc: 0.9844 - val_loss: 0.0277 - val_acc: 0.9913

Epoch 5

60000/60000 [==============================] - 79s - loss: 0.0467 - acc: 0.9859 - val_loss: 0.0305 - val_acc: 0.9895

Epoch 6

60000/60000 [==============================] - 79s - loss: 0.0416 - acc: 0.9868 - val_loss: 0.0291 - val_acc: 0.9911

Epoch 7

60000/60000 [==============================] - 79s - loss: 0.0377 - acc: 0.9882 - val_loss: 0.0289 - val_acc: 0.9909

Epoch 8

60000/60000 [==============================] - 79s - loss: 0.0349 - acc: 0.9889 - val_loss: 0.0263 - val_acc: 0.9914

Epoch 9

60000/60000 [==============================] - 79s - loss: 0.0333 - acc: 0.9900 - val_loss: 0.0258 - val_acc: 0.9916

Epoch 10

60000/60000 [==============================] - 79s - loss: 0.0305 - acc: 0.9906 - val_loss: 0.0240 - val_acc: 0.9924

Epoch 11

60000/60000 [==============================] - 79s - loss: 0.0273 - acc: 0.9918 - val_loss: 0.0299 - val_acc: 0.9906

Test score: 0.0299028589356

Test accuracy: 0.9906