TensorFlow模型加载与保存

我们经常遇到训练时间很长，使用起来就是Weight和Bias。那么如何将训练和测试分开操作呢？

TF给出了模型的加载与保存操作，看了网上都是很简单的使用了一下，这里给出一个神经网络的小程序去测试。

本博文使用了Titanic的数据进行操作：

Train.Py

import numpy as np
import pandas as pd
import tensorflow as tf
from sklearn.model_selection import train_test_split

################################
# Preparing Data
################################

# read data from file
data = pd.read_csv('data/train.csv')

# fill nan values with 0
data = data.fillna(0)
# convert ['male', 'female'] values of Sex to [1, 0]
data['Sex'] = data['Sex'].apply(lambda s: 1 if s == 'male' else 0)
# 'Survived' is the label of one class,
# add 'Deceased' as the other class
data['Deceased'] = data['Survived'].apply(lambda s: 1 - s)

# select features and labels for training
dataset_X = data[['Sex', 'Age', 'Pclass', 'SibSp', 'Parch', 'Fare']].as_matrix()
dataset_Y = data[['Deceased', 'Survived']].as_matrix()

# split training data and validation set data
X_train, X_val, y_train, y_val = train_test_split(dataset_X, dataset_Y,
                                                  test_size=0.2,
                                                  random_state=42)

################################
# Constructing Dataflow Graph
################################

# create symbolic variables
X = tf.placeholder(tf.float32, shape=[None, 6])
y = tf.placeholder(tf.float32, shape=[None, 2])

# weights and bias are the variables to be trained
weights = tf.Variable(tf.random_normal([6, 2]), name='weights')
bias = tf.Variable(tf.zeros([2]), name='bias')
y_pred = tf.nn.softmax(tf.matmul(X, weights) + bias)

# Minimise cost using cross entropy
# NOTE: add a epsilon(1e-10) when calculate log(y_pred),
# otherwise the result will be -inf
cross_entropy = - tf.reduce_sum(y * tf.log(y_pred + 1e-10),
                                reduction_indices=1)
cost = tf.reduce_mean(cross_entropy)

# use gradient descent optimizer to minimize cost
train_op = tf.train.GradientDescentOptimizer(0.001).minimize(cost)

# calculate accuracy
correct_pred = tf.equal(tf.argmax(y, 1), tf.argmax(y_pred, 1))
acc_op = tf.reduce_mean(tf.cast(correct_pred, tf.float32))

################################
# Training and Evaluating the model
################################
saver = tf.train.Saver()
# use session to run the calculation
with tf.Session() as sess:
    # variables have to be initialized at the first place
    tf.global_variables_initializer().run()
    # training loop
    for epoch in range(10):
        total_loss = 0.
        for i in range(len(X_train)):
            # prepare feed data and run
            feed_dict = {X: [X_train[i]], y: [y_train[i]]}
            _, loss = sess.run([train_op, cost], feed_dict=feed_dict)
            total_loss += loss
        # display loss per epoch
        print('Epoch: %04d, total loss=%.9f' % (epoch + 1, total_loss))
    saver_path = saver.save(sess,"wjy_data/model.ckpt")
    # Accuracy calculated by TensorFlow
    accuracy = sess.run(acc_op, feed_dict={X: X_val, y: y_val})
    print("Accuracy on validation set: %.9f" % accuracy)

    # Accuracy calculated by NumPy
    pred = sess.run(y_pred, feed_dict={X: X_val})
    correct = np.equal(np.argmax(pred, 1), np.argmax(y_val, 1))
    numpy_accuracy = np.mean(correct.astype(np.float32))
    print("Accuracy on validation set (numpy): %.9f" % numpy_accuracy)

    # predict on test data
    testdata = pd.read_csv('data/test.csv')
    testdata = testdata.fillna(0)
    # convert ['male', 'female'] values of Sex to [1, 0]
    testdata['Sex'] = testdata['Sex'].apply(lambda s: 1 if s == 'male' else 0)
    X_test = testdata[['Sex', 'Age', 'Pclass', 'SibSp', 'Parch', 'Fare']]
    predictions = np.argmax(sess.run(y_pred, feed_dict={X: X_test}), 1)
    submission = pd.DataFrame({
        "PassengerId": testdata["PassengerId"],
        "Survived": predictions
    })

    submission.to_csv("titanic-submission.csv", index=False)

注意：

　　saver_path = saver.save(sess,"wjy_data/model.ckpt")
　　项目目录下面必须新建一个wjy_data的文件夹，不然会报错！！！

Test.Py

import numpy as np
import pandas as pd
import tensorflow as tf
from sklearn.model_selection import train_test_split

# create symbolic variables
X = tf.placeholder(tf.float32, shape=[None, 6])
y = tf.placeholder(tf.float32, shape=[None, 2])

# weights and bias are the variables to be trained
weights = tf.Variable(tf.random_normal([6, 2]), name='weights')
bias = tf.Variable(tf.zeros([2]), name='bias')
y_pred = tf.nn.softmax(tf.matmul(X, weights) + bias)

# predict on test data
testdata = pd.read_csv('data/test.csv')
testdata = testdata.fillna(0)
# convert ['male', 'female'] values of Sex to [1, 0]
testdata['Sex'] = testdata['Sex'].apply(lambda s: 1 if s == 'male' else 0)
X_test = testdata[['Sex', 'Age', 'Pclass', 'SibSp', 'Parch', 'Fare']]
################################
# Training and Evaluating the model
################################
saver = tf.train.Saver()
# use session to run the calculation
with tf.Session() as sess:
    # variables have to be initialized at the first place
    tf.global_variables_initializer().run()
    #save_path = saver.save(sess,"Saved_model/model.ckpt")
    saver.restore(sess,"wjy_data/model.ckpt")#加载模型
    predictions = np.argmax(sess.run(y_pred, feed_dict={X: X_test}), 1)
    submission = pd.DataFrame({
        "PassengerId": testdata["PassengerId"],
        "Survived": predictions
    })
    #saver = tf.train.Saver()
    submission.to_csv("titanic-submission.csv", index=False)

很方便的使用保存模型的方式去测试和训练数据，不然怎么办~~

参考：

　　《深度学习原理与TensorFlow实战》

　　https://blog.csdn.net/lujiandong1/article/details/53301994