大家都知道,在加载模型及对应的权重进行训练的时候,我们可以整个使用所提供的模型,但是有时候呢?所提供的模型不能很好的满足我们的要求,有时候我们只需要模型的前几层然后进行对应的权重赋值,这时候,我们应该怎么办呢?tensorflow为我们提供了两种方法(探索了好久才找到解决办法,不过感觉蛮有用的,分享给大家啦!):
1)在加载模型的时候,使用final_endpoint参数,指定模型阶段点:
import tensorflow.contrib.slim.nets as nets
#from tensorflow.contrib.slim.nets.inception import inception_v3, inception_v3_arg_scope
import numpy as np
import os
height = 299
width = 299
channels = 3
num_classes=1001
X = tf.placeholder(tf.float32, shape=[None, height, width, channels])
y = tf.placeholder(tf.float32,shape=[None,182])
with slim.arg_scope(nets.inception.inception_v3_arg_scope()):
logits, end_points = nets.inception.inception_v3_base(X,final_endpoint = "Mixed_7a")
variables_to_restore=slim.get_variables_to_restore()
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
saver=tf.train.Saver(variables_to_restore)
saver.restore(sess,os.path.join("E:\\","inception_v3.ckpt"))
print("Done")
2)在恢复模型的时候,部分层不使用ckpt文件中提供的参数:
#-*-coding=utf-8-*-
import tensorflow as tf
import tensorflow.contrib.slim as slim
import tensorflow.contrib.slim.nets as nets
#from tensorflow.contrib.slim.nets.inception import inception_v3, inception_v3_arg_scope
import numpy as np
import os
height = 299
width = 299
channels = 3
num_classes=1001
X = tf.placeholder(tf.float32, shape=[None, height, width, channels])
y = tf.placeholder(tf.float32,shape=[None,182])
with slim.arg_scope(nets.inception.inception_v3_arg_scope()):
logits, end_points = nets.inception.inception_v3(X, num_classes=num_classes,is_training=False)
with tf.Session() as sess:
exclude=['Mixed_7c','Mixed_7b','AuxLogits','AuxLogits','Logits','Predictions']
variables_to_restore=slim.get_variables_to_restore(exclude=exclude)
sess.run(tf.global_variables_initializer())
saver=tf.train.Saver(variables_to_restore)
saver.restore(sess,os.path.join("E:\\","inception_v3.ckpt"))
print("Done")