一、模型文件简介
Tensorflow模型主要包含网络参数的网络设计或图，以及训练的网络参数的值，因此包含两个文件
a)
meta graph--存储
Tensorflow图形;即所有的变量，操作，集合等等。这个文件有扩展文件.meta。
b)
checkpoint --一个二进制文件，包含了所有的权重、偏差、梯度和所有其他保存的变量的值。这个文件有一个扩展文件.ckpt。然而，Tensorflow已经从0.11版本中改变了这一点。现在，我们有两个文件，而不是单独的ckpt文件。
mymodel.data-00000-of-00001 --包含训练的变量
mymodel.index
所以大于0.1版本的模型文件将有4个
0.11版本以前是三个文件
checkpoint
checkpoint
inception_v1.meta
inception_v1.ckpt
二、存储模型save
1，由于tf中变量只能存在会话session中，所以必须在会话中存储
import tensorflow as tf
w1 = tf.Variable(tf.random_normal(shape=[2]), name='w1')
w2 = tf.Variable(tf.random_normal(shape=[5]), name='w2')
saver = tf.train.Saver()
sess = tf.Session()
sess.run(tf.global_variables_initializer())
saver.save(sess, 'my_test_model')
# This will save following files in Tensorflow v >= 0.11
# my_test_model.data-00000-of-00001
# my_test_model.index
# my_test_model.meta
# checkpoint,2，每1000步迭代进行一次存储
saver.save(sess, 'my_test_model',global_step=1000)
存储文件结构
my_test_model-1000.index
my_test_model-1000.meta
my_test_model-1000.data-00000-of-00001
checkpoint
3，由于meta文件不会改变，如果不想每次迭代都存储该文件（即只存储一个.meta文件）
saver.save(sess, 'my-model', global_step=step,write_meta_graph=False)4，如果您想只保留4个最新模型，并希望在训练后每2小时保存一个模型，那么您可以使用max_to_keep和keep_checkpoint_every_n_hours。
#saves a model every 2 hours and maximum 4 latest models are saved.
saver = tf.train.Saver(max_to_keep=4, keep_checkpoint_every_n_hours=2)5，指定存储某些变量或者集合（variable/collection)
import tensorflow as tf
w1 = tf.Variable(tf.random_normal(shape=[2]), name='w1')
w2 = tf.Variable(tf.random_normal(shape=[5]), name='w2')
saver = tf.train.Saver([w1,w2])
sess = tf.Session()
sess.run(tf.global_variables_initializer())
saver.save(sess, 'my_test_model',global_step=1000)
三、导入预训练模型restore
1，使用预训练模型进行微调
a)创建网络，可以自己手动写。也可以使用.meta文件重建网络，但记得还要加载参数
saver = tf.train.import_meta_graph('my_test_model-1000.meta')b)加载参数
with tf.Session() as sess:
new_saver = tf.train.import_meta_graph('my_test_model-1000.meta')
new_saver.restore(sess, tf.train.latest_checkpoint('./'))2，对于只存储了某些变量或者集合的情况，如存储了w1和w2
with tf.Session() as sess:
saver = tf.train.import_meta_graph('my-model-1000.meta')
saver.restore(sess,tf.train.latest_checkpoint('./'))
print(sess.run('w1:0'))
##Model has been restored. Above statement will print the saved value of w1.
四、恢复任何预先训练过的模型，并将其用于预测，微调或进一步训练
1，当您使用Tensorflow时，您可以定义一个图表喂训练数据和一些超参数，比如学习速率，全局步骤等等。使用占位符来输入所有训练数据和超参数的标准做法。让我们用占位符构建一个小网络，并保存它。注意，当网络被保存时，占位符的值不会被保存。
import tensorflow as tf
#Prepare to feed input, i.e. feed_dict and placeholders
w1 = tf.placeholder("float", name="w1")
w2 = tf.placeholder("float", name="w2")
b1= tf.Variable(2.0,name="bias")
feed_dict ={w1:4,w2:8}
#Define a test operation that we will restore
w3 = tf.add(w1,w2)
w4 = tf.multiply(w3,b1,name="op_to_restore")
sess = tf.Session()
sess.run(tf.global_variables_initializer())
#Create a saver object which will save all the variables
saver = tf.train.Saver()
#Run the operation by feeding input
print(sess.run(w4,feed_dict))
#Prints 24 which is sum of (w1+w2)*b1
#Now, save the graph
saver.save(sess, 'my_test_model',global_step=1000)2，我们可以通过graph.get_tensor_by_name()方法获得这些保存的操作和占位符变量的引用。
#How to access saved variable/Tensor/placeholders
w1 = graph.get_tensor_by_name("w1:0")
## How to access saved operation
op_to_restore = graph.get_tensor_by_name("op_to_restore:0")3，如果只想用不同数据跑相同网络，通过feed_dict喂新数据就行
import tensorflow as tf
sess=tf.Session()
#First let's load meta graph and restore weights
saver = tf.train.import_meta_graph('my_test_model-1000.meta')
saver.restore(sess,tf.train.latest_checkpoint('./'))
# Now, let's access and create placeholders variables and
# create feed-dict to feed new data
graph = tf.get_default_graph()
w1 = graph.get_tensor_by_name("w1:0")
w2 = graph.get_tensor_by_name("w2:0")
feed_dict ={w1:13.0,w2:17.0}
#Now, access the op that you want to run.
op_to_restore = graph.get_tensor_by_name("op_to_restore:0")
print(sess.run(op_to_restore,feed_dict))
#This will print 60 which is calculated
#using new values of w1 and w2 and saved value of b1. 4， 如果想添加更多操作到图里
import tensorflow as tf
sess=tf.Session()
#First let's load meta graph and restore weights
saver = tf.train.import_meta_graph('my_test_model-1000.meta')
saver.restore(sess,tf.train.latest_checkpoint('./'))
# Now, let's access and create placeholders variables and
# create feed-dict to feed new data
graph = tf.get_default_graph()
w1 = graph.get_tensor_by_name("w1:0")
w2 = graph.get_tensor_by_name("w2:0")
feed_dict ={w1:13.0,w2:17.0}
#Now, access the op that you want to run.
op_to_restore = graph.get_tensor_by_name("op_to_restore:0")
#Add more to the current graph
add_on_op = tf.multiply(op_to_restore,2)
print(sess.run(add_on_op,feed_dict))
#This will print 120.5，恢复旧图的一部分功能进行微调的，使用graph.get_tensor_by_name()方法访问适当的操作，并在此基础上构建图表。这是一个真实的例子。在这里，我们使用元图加载一个vgg预先训练的网络，并在最后一层将输出的数量更改为2，以便对新数据进行微调。
......
......
saver = tf.train.import_meta_graph('vgg.meta')
# Access the graph
graph = tf.get_default_graph()
## Prepare the feed_dict for feeding data for fine-tuning
#Access the appropriate output for fine-tuning
fc7= graph.get_tensor_by_name('fc7:0')
#use this if you only want to change gradients of the last layer
fc7 = tf.stop_gradient(fc7) # It's an identity function
fc7_shape= fc7.get_shape().as_list()
new_outputs=2
weights = tf.Variable(tf.truncated_normal([fc7_shape[3], num_outputs], stddev=0.05))
biases = tf.Variable(tf.constant(0.05, shape=[num_outputs]))
output = tf.matmul(fc7, weights) + biases
pred = tf.nn.softmax(output)
# Now, you run this with fine-tuning data in sess.run()来自官方文档http://cv-tricks.com/tensorflow-tutorial/save-restore-tensorflow-models-quick-complete-tutorial/