如何在 ipython 或 anaconda 中恢复 TensorFlow 中的检查点

How to restore checkpoint in TensorFlow inside ipython or anaconda

我用的是tensorflow 0.9.
我想保存我的模型然后恢复它。
我只是添加 tf.train.Saver() 来保存和恢复我的训练变量。

这是我的代码:

import tensorflow as tf
import input_data
import os

checkpoint_dir='./ckpt_dir/'

mnist = input_data.read_data_sets("MNIST_data", one_hot = True)

x = tf.placeholder(tf.float32, shape = [None , 784])
y_ = tf.placeholder(tf.float32, [None, 10])

sess = tf.InteractiveSession()

def load_model(sess, saver, checkpoint_dir ):

ckpt = tf.train.get_checkpoint_state(checkpoint_dir)
if ckpt and ckpt.model_checkpoint_path:
print(ckpt.model_checkpoint_path)

saver.restore(sess, ckpt.model_checkpoint_path)

else:
if not os.path.exists(checkpoint_dir):
os.makedirs(checkpoint_dir)
sess.run(init)
return

def weight_variable(shape):
initial = tf.truncated_normal(shape, stddev = 0.1)
return tf.Variable(initial)

def bias_variable(shape):
initial = tf.constant(0.1, shape= shape)
return tf.Variable(initial)

def conv2d(x, W):
return tf.nn.conv2d(x, W, strides = [1, 1, 1, 1], padding = "SAME")

def max_pool_2x2(x):
return tf.nn.max_pool(x, ksize = [1, 2, 2, 1], strides = [1, 2, 2, 1],
padding = "SAME")

W_conv1 = weight_variable([5, 5, 1, 32])
b_conv1 = bias_variable([32])

x_image = tf.reshape(x, [-1, 28, 28, 1])

#
h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1))
h_pool1 = max_pool_2x2(h_conv1)

#
W_conv2 = weight_variable([5, 5, 32, 64])
b_conv2 = bias_variable([64])

h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2))
h_pool2 = max_pool_2x2(h_conv2)

W_fc1 = weight_variable([7764, 1024])
b_fc1 = bias_variable([1024])

h_pool2_flat = tf.reshape(h_pool2, [-1, 7764])
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)

#
keep_prob = tf.placeholder(tf.float32)
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)

#
W_fc2 = weight_variable([1024, 10])
b_fc2 = bias_variable([10])

y_conv = tf.nn.softmax(tf.matmul(h_fc1_drop,W_fc2) +b_fc2)

#
cross_entropy = tf.reduce_mean(-tf.reduce_sum(y_ * tf.log(y_conv), reduction_indices = [1]))
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)

correct_prediction = tf.equal(tf.argmax(y_conv, 1), tf.argmax(y_, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

init = tf.initialize_all_variables()

saver = tf.train.Saver()

load_model(sess, saver, checkpoint_dir)

for i in range(1):
batch = mnist.train.next_batch(50)
if i%10 == 0:
train_accuracy = accuracy.eval(feed_dict = {x : batch[0] , y_ : batch[1], keep_prob : 1.0})
print("step %d, training accuracy %g"%(i, train_accuracy))

train_step.run(feed_dict = {x : batch[0], y_ : batch[1], keep_prob : 0.5})
print("test accuracy %g"%accuracy.eval(feed_dict={
x: mnist.test.images, y_: mnist.test.labels, keep_prob: 1.0}))

tf.scalar_summary("accuracy", accuracy)

saver.save(sess,checkpoint_dir+'model.ckpt')

当我恢复检查点时:

saver.restore(sess, ckpt.model_checkpoint_path)

TensorFlow 抛出此错误:

Traceback (most recent call last):
.
.
.
NotFoundError: Tensor name "global_step_7" not found in checkpoint files ./ckpt_dir/model.ckpt-0
[[Node: save_18/restore_slice_438 = RestoreSlicedt=DT_INT32, preferred_shard=-1, _device="/job:localhost/replica:0/task:0/cpu:0"]]
Caused by op 'save_18/restore_slice_438', defined at:
File "/home/m/anaconda3/lib/python3.5/site-packages/spyderlib/widgets/externalshell/start_ipython_kernel.py", line 205, in
ipythonkernel.start()
.
.
.
File "/home/m/anaconda3/lib/python3.5/site-packages/tensorflow/python/framework/ops.py", line 1224, in __init
raise TypeError("Control input must be an Operation, "

编辑:

我用蟒蛇。我第一次在 spyder 或 ipython 中 运行 这段代码使用“运行 filename.py”,它将模型保存在检查点中,但是当我 运行 这个再次代码,它会抛出错误。

但是当我关闭 spyder 或 ipython 时,再次打开它并且 运行 它的代码正确地恢复了检查点。

另外,当我在终端“python filename.py”中 运行 时,它总是 运行 并且不会抛出任何错误。

当您再次 运行 文件时,您需要在通话开始时重置默认图表。


如果你不重置默认图形,并且运行两倍行:

x = tf.Variable(1, name='x')
print x.name

你会看到第一次 x 的名字是 "x:0",第二次是 "x_1:0"。这就是让 tf.train.Saver:

感到困惑的地方
  • 它首先使用名称 "x:0"
  • 保存 x 的值
  • 然后在接下来的运行中你尝试加载x的保存值,但是现在变量的名称是"x_1:0",所以saver尝试加载一个保存的值名称 "x_1:0" 下的值但找不到它,并且 returns 出错。

但是,您可以使用 tf.reset_default_graph() 在开始时重置默认图表。这将创建一个空图并将其用作默认图。
这里 x 的名称在这两个图中可以相同:

# First run
tf.reset_default_graph()
x = tf.Variable(1, name='x')
print x.name  # prints 'x:0'

# Next run
tf.reset_default_graph()
x = tf.Variable(1, name='x')
print x.name  # prints 'x:0'

这两个变量现在可以具有相同的名称,因为它们不再在同一个图中。


另一种方法是在开始时创建一个图形并将其用作默认图形:

graph = tf.Graph()
with graph.as_default():
    x = tf.Variable(1, name='x')