from __future__ import division,print_function,absolute_import
import tflearn
from tflearn.data_utils import shuffle,to_categorical
from tflearn.layers.core import input_data,dropout,fully_connected
from tflearn.layers.conv import conv_2d,max_pool_2d
from tflearn.layers.estimator import regression
from tflearn.data_preprocessing import ImagePreprocessing
from tflearn.data_augmentation import ImageAugmentation
from tflearn.datasets import cifar10
#import data
(X,Y),(X_test,Y_test)=cifar10.load_data(dirname="cifar10") #load_data数据也可以为空，自动下载，比较耗时
X,Y=shuffle(X,Y)
Y=to_categorical(Y,10)
Y_test=to_categorical(Y_test,10)
#preprocess
img_prep=ImagePreprocessing()
img_prep.add_featurewise_zero_center()
img_prep.add_featurewise_stdnorm()
#augmentation
img_aug=ImageAugmentation()
img_aug.add_random_flip_leftright()
img_aug.add_random_rotation(max_angle=25.)
#define networks
network=input_data(shape=[None,32,32,3],data_preprocessing=img_prep,data_augmentation=img_aug)
network=conv_2d(network,32,3,activation='relu')
network=max_pool_2d(network,2)
network=conv_2d(network,64,3,activation='relu')
network=conv_2d(network,64,3,activation='relu')
network=max_pool_2d(network,2)
network=fully_connected(network,512,activation='relu')
network=dropout(network,0.5)
network=fully_connected(network,10,activation='softmax')
network=regression(network,optimizer='adam',loss='categorical_crossentropy',learning_rate=0.001)
#train
model=tflearn.DNN(network,tensorboard_verbose=0)
model.fit(X,Y,n_epoch=50,shuffle=True,validation_set=(X_test,Y_test),
show_metric=True,batch_size=96,run_id='cifar10_cnn')
-------------------------------------------------
Training Step: 5265 | total loss: 0.64463 | time: 0.558s
| Adam | epoch: 011 | loss: 0.64463 - acc: 0.7726 -- iter: 05280/50000
Training Step: 5266 | total loss: 0.66460 | time: 0.568s
| Adam | epoch: 011 | loss: 0.66460 - acc: 0.7672 -- iter: 05376/50000
Training Step: 5267 | total loss: 0.66508 | time: 0.579s
| Adam | epoch: 011 | loss: 0.66508 - acc: 0.7665 -- iter: 05472/50000
Training Step: 5268 | total loss: 0.65549 | time: 0.598s
| Adam | epoch: 011 | loss: 0.65549 - acc: 0.7722 -- iter: 05568/50000
在epoch11次的时候，acc为0.772，可以多迭代几次，精度会更高
