Tensor is not an element of this graph tensorflow/Failed to create session.

最新推荐文章于 2023-04-06 02:13:12 发布

原创最新推荐文章于 2023-04-06 02:13:12 发布 · 268 阅读

1 ·

CC 4.0 BY-SA版权

文章标签：

#Tensorflow

本文通过对比两种不同的代码实现方式，详细解析了在使用TensorFlow时遇到的Tensor is not an element of this graph及Failed to create session错误原因。并提供了正确的解决方案，确保所有操作都在同一个图中执行。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

Tensor is not an element of this graph tensorflow / Failed to create session.

在运行Tensorflow时，遇到Tensor is not an element of this graph tensorflow/Failed to create session问题。基于 https://blog.youkuaiyun.com/lujiandong1/article/details/53448012 博主博客的基础上，进行扩展。

import tensorflow as tf
def activation(e, f, g):
    return e + f + g
with tf.Graph().as_default():
    a = tf.constant([5, 4, 5], name='a')
    b = tf.constant([0, 1, 2], name='b')
    c = tf.constant([5, 0, 5], name='c')
    res = activation(a, b, c)
    init = tf.initialize_all_variables()
with tf.Session() as sess:    # Start running operations on the Graph.
    sess.run(init)
    hi = sess.run(res)
    print (hi)

这样运行会报错：
tensorflow.python.framework.errors_impl.InternalError: Failed to create session.
因为，with tf.Session() as sess: 与with tf.Graph().as_default()：相当于这sess是新建了一个图，与with tf.Graph().as_default()不是一个图。但是去掉with tf.Graph().as_default():后：

import tensorflow as tf
def activation(e, f, g):
    return e + f + g
a = tf.constant([5, 4, 5], name='a')
b = tf.constant([0, 1, 2], name='b')
c = tf.constant([5, 0, 5], name='c')
res = activation(a, b, c)
init = tf.initialize_all_variables()
with tf.Session() as sess:    # Start running operations on the Graph.
    sess.run(init)
    hi = sess.run(res)
    print (hi)

这样会正常运行，因此这样所有的op都只会在默认的graph存在。