[机器学习]RuntimeError: The Session graph is empty. Add operations to the graph before calling run().

在进行机器学习的过程中

 1 #!/usr/bin/env python
 2 # -*- coding: utf-8 -*-
 3 # @File  : TensorFlow入门操作.py
 4 # @Author: zhao
 5 # @Date  : 2020/3/26
 6 # @Desc  :
 7 # @Contact : 398333404@qq.com 
 8 
 9 import tensorflow as tf
10 import os
11 os.environ["CUDA_VISIBLE_DEVICES"]="0"
12 
13 
14 hello=tf.constant('Hello,TensorFlow')
15 config=tf.compat.v1.ConfigProto()
16 config.gpu_options.per_process_gpu_memory_fraction = 0.9
17 sess=tf.compat.v1.Session(config=config)
18 print(sess.run(hello))

会报如下错误

　　raise RuntimeError('The Session graph is empty. Add operations to the '
RuntimeError: The Session graph is empty. Add operations to the graph before calling run().

• 问题产生的原因：无法执行sess.run()的原因是tensorflow版本不同导致的，tensorflow版本2.0无法兼容版本1.0.
• 解决办法：
•  tf.compat.v1.disable_eager_execution() 

修改后的代码

 1 #!/usr/bin/env python
 2 # -*- coding: utf-8 -*-
 3 # @File  : TensorFlow入门操作.py
 4 # @Author: 赵路仓
 5 # @Date  : 2020/3/26
 6 # @Desc  :
 7 # @Contact : 398333404@qq.com 
 8 
 9 import tensorflow as tf
10 import os
11 os.environ["CUDA_VISIBLE_DEVICES"]="0"
12 
13 tf.compat.v1.disable_eager_execution()
14 hello=tf.constant('Hello,TensorFlow')
15 config=tf.compat.v1.ConfigProto()
16 config.gpu_options.per_process_gpu_memory_fraction = 0.9
17 sess=tf.compat.v1.Session(config=config)
18 print(sess.run(hello))