本项目所用到的软件及库：
1. Tensorflow
2. keras
3. cv2
4. sklearn
5. numpy
6. pandas
7. h5py

本项目数据集地址：
http://www.kaggle.com/c/dogs-vs-cats

本项目运行所需时间：p2-xlarge上约1小时，主要时间是从预训练模型导出特征比较耗时。

附： 最终模型训练结果：
Train on 19973 samples, validate on 4994 samples
Epoch 1/64
19973/19973 [==============================] - 5s 226us/step - loss: 0.1958 - acc: 0.9293 - val_loss: 0.0494 - val_acc: 0.9914
Epoch 2/64
19973/19973 [==============================] - 4s 190us/step - loss: 0.0453 - acc: 0.9903 - val_loss: 0.0300 - val_acc: 0.9920
Epoch 3/64
19973/19973 [==============================] - 4s 193us/step - loss: 0.0313 - acc: 0.9926 - val_loss: 0.0235 - val_acc: 0.9948
Epoch 4/64
19973/19973 [==============================] - 4s 182us/step - loss: 0.0252 - acc: 0.9938 - val_loss: 0.0202 - val_acc: 0.9952
Epoch 5/64
19973/19973 [==============================] - 4s 190us/step - loss: 0.0214 - acc: 0.9936 - val_loss: 0.0184 - val_acc: 0.9954
Epoch 6/64
19973/19973 [==============================] - 4s 181us/step - loss: 0.0191 - acc: 0.9943 - val_loss: 0.0172 - val_acc: 0.9952
Epoch 7/64
19973/19973 [==============================] - 4s 183us/step - loss: 0.0183 - acc: 0.9945 - val_loss: 0.0162 - val_acc: 0.9954
Epoch 8/64
19973/19973 [==============================] - 4s 189us/step - loss: 0.0169 - acc: 0.9949 - val_loss: 0.0155 - val_acc: 0.9956
Epoch 9/64
19973/19973 [==============================] - 4s 198us/step - loss: 0.0162 - acc: 0.9947 - val_loss: 0.0151 - val_acc: 0.9960
Epoch 10/64
19973/19973 [==============================] - 4s 207us/step - loss: 0.0151 - acc: 0.9954 - val_loss: 0.0148 - val_acc: 0.9956
Epoch 11/64
19973/19973 [==============================] - 4s 212us/step - loss: 0.0136 - acc: 0.9959 - val_loss: 0.0145 - val_acc: 0.9954
Epoch 12/64
19973/19973 [==============================] - 4s 205us/step - loss: 0.0137 - acc: 0.9953 - val_loss: 0.0142 - val_acc: 0.9954
Epoch 13/64
19973/19973 [==============================] - 4s 206us/step - loss: 0.0131 - acc: 0.9960 - val_loss: 0.0141 - val_acc: 0.9962
Epoch 14/64
19973/19973 [==============================] - 4s 208us/step - loss: 0.0127 - acc: 0.9956 - val_loss: 0.0139 - val_acc: 0.9962
Epoch 15/64
19973/19973 [==============================] - 4s 217us/step - loss: 0.0119 - acc: 0.9959 - val_loss: 0.0137 - val_acc: 0.9962
Epoch 16/64
19973/19973 [==============================] - 4s 207us/step - loss: 0.0116 - acc: 0.9966 - val_loss: 0.0136 - val_acc: 0.9960
Epoch 17/64
19973/19973 [==============================] - 4s 202us/step - loss: 0.0103 - acc: 0.9969 - val_loss: 0.0136 - val_acc: 0.9960
Epoch 18/64
19973/19973 [==============================] - 4s 213us/step - loss: 0.0106 - acc: 0.9969 - val_loss: 0.0134 - val_acc: 0.9962
Epoch 19/64
19973/19973 [==============================] - 4s 214us/step - loss: 0.0099 - acc: 0.9968 - val_loss: 0.0134 - val_acc: 0.9958
Epoch 20/64
19973/19973 [==============================] - 4s 203us/step - loss: 0.0099 - acc: 0.9967 - val_loss: 0.0132 - val_acc: 0.9960
Epoch 21/64
19973/19973 [==============================] - 4s 208us/step - loss: 0.0103 - acc: 0.9968 - val_loss: 0.0132 - val_acc: 0.9960
Epoch 22/64
19973/19973 [==============================] - 4s 206us/step - loss: 0.0095 - acc: 0.9969 - val_loss: 0.0133 - val_acc: 0.9960
Epoch 23/64
19973/19973 [==============================] - 4s 206us/step - loss: 0.0096 - acc: 0.9972 - val_loss: 0.0130 - val_acc: 0.9964
Epoch 24/64
19973/19973 [==============================] - 4s 207us/step - loss: 0.0090 - acc: 0.9969 - val_loss: 0.0132 - val_acc: 0.9964
Epoch 25/64
19973/19973 [==============================] - 4s 208us/step - loss: 0.0081 - acc: 0.9976 - val_loss: 0.0132 - val_acc: 0.9964
Epoch 26/64
19973/19973 [==============================] - 4s 207us/step - loss: 0.0087 - acc: 0.9972 - val_loss: 0.0132 - val_acc: 0.9962
12500/12500 [==============================] - 1s 58us/step
Found 12500 images belonging to 1 classes.