在TensorFlow中修改已恢复的CNN模型的权重和偏差

时间:2017-12-09 00:04:20

标签: python machine-learning tensorflow deep-learning

我最近开始使用TensorFlow(TF),我遇到了一个需要帮助的问题。基本上,我已经恢复了预先训练的模型,在重新测试其准确性之前,我需要修改其中一个层的权重和偏差。现在,我的问题如下: 如何使用TF中的assign方法更改权重和偏差?是否可以在TF中修改已恢复建模的权重?

这是我的代码:

import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data # Imports the MINST dataset

# Data Set:
# ---------
mnist = input_data.read_data_sets("/home/frr/MNIST_data", one_hot=True)# An object where data is stored

ImVecDim = 784# The number of elements in a an image vector (flattening a 28x28 2D image)
NumOfClasses = 10

g = tf.get_default_graph()

with tf.Session() as sess:
  LoadMod = tf.train.import_meta_graph('simple_mnist.ckpt.meta')  # This object loads the model
  LoadMod.restore(sess, tf.train.latest_checkpoint('./'))# Loading weights and biases and other stuff to the model

  # ( Here I'd like to modify the weights and biases of layer 1, set them to one for example, before I go ahead and test the accuracy ) #

  # Testing the acuracy of the model:
  X = g.get_tensor_by_name('ImageIn:0')
  Y = g.get_tensor_by_name('LabelIn:0')
  KP = g.get_tensor_by_name('KeepProb:0')
  Accuracy = g.get_tensor_by_name('NetAccuracy:0')
  feed_dict = { X: mnist.test.images[:256], Y: mnist.test.labels[:256], KP: 1.0 }
  print( 'Model Accuracy = ' )
  print( sess.run( Accuracy, feed_dict ) )

3 个答案:

答案 0 :(得分:1)

除现有答案外,还可以通过tf.assign函数执行张量更新。

v1 = sess.graph.get_tensor_by_name('v1:0')
print(sess.run(v1))   # 1.0
sess.run(tf.assign(v1, v1 + 1))
print(sess.run(v1))   # 2.0

答案 1 :(得分:0)

是的,这是可能的。加载元图后,您的权重和偏差已经加载。您需要找到它们的名称(请参阅list_variables函数),然后将它们分配给Python变量。

为此,请将time(nullptr)与变量名称一起使用。您可能必须在变量范围上设置tf.get_variable。有关重用变量的详细信息,请参阅this answer

将它们作为reuse=True变量后,您可以拨打weights

答案 2 :(得分:0)

感谢所有回复的人。我只想把各个部分拼凑起来。这是帮助我完成我想要的代码:

db.collection(collection).add(documentData)
.then(ref => {
    console.log("Document written with ID: ", ref.id);
});