entityMap|0|type|LINK|mutability|MUTABLE|data|url|https://www.tensorflow.org/guide/eager|blocks|key|2kosb|text|tf.placeholder()被提供给会话，在运行时接收来自提要dict的值并执行所需的操作。通常，您会创建一个带有'+with‘关键字的Session()并运行它。但这可能并不有利于所有需要立即执行的情况。这就是所谓的急切执行。示例：|unstyled|depth|inlineStyleRanges|entityRanges|8s8kq|通常，这是运行会话的过程：|8hpuc|import+tensorflow+as+tf

def+square(num):
++++return+tf.square(num)+

p+=+tf.placeholder(tf.float32)
q+=+square(num)

with+tf.Session()+as+sess:
++++print(sess.run(q,+feed_dict={num:+10})|code-block|syntax|javascript|dm3hh|但是，当我们以急切的执行方式运行时，我们是这样运行的：|d0u9h|import+tensorflow+as+tf

tf.enable_eager_execution()

def+square(num):
+++return+tf.square(num)

print(square(10))+|1anpe|因此，我们不需要在会话中显式地运行它，并且在大多数情况下可以更直观。这提供了更多的交互式执行。欲了解更多详情，请访问：https://www.tensorflow.org/guide/eager|offset|length|4eapt|如果您要将代码从tensorflow+v1转换为tensorflow+v2，则必须实现tf.compat.v1，并且占位符出现在tf.compat.v1.placeholder中，但这只能在关闭急切模式下执行。|2mpfi|tf.compat.v1.disable_eager_execution()|fir7p|TensorFlow发布了急切执行模式，每个节点在定义后都会立即执行。因此，使用tf.placeholder的语句不再有效。|6good^0|0|0|0|0|0|1N|12|0|0|0|0|0^^$0|$1|$2|3|4|5|6|$7|8]]]|9|@$A|B|C|D|2|E|F|14|G|@]|H|@]|6|$]]|$A|I|C|J|2|E|F|15|G|@]|H|@]|6|$]]|$A|K|C|L|2|M|F|16|G|@]|H|@]|6|$N|O]]|$A|P|C|Q|2|E|F|17|G|@]|H|@]|6|$]]|$A|R|C|S|2|M|F|18|G|@]|H|@]|6|$N|O]]|$A|T|C|U|2|E|F|19|G|@]|H|@$V|1A|W|1B|A|1C]]|6|$]]|$A|X|C|Y|2|E|F|1D|G|@]|H|@]|6|$]]|$A|Z|C|10|2|M|F|1E|G|@]|H|@]|6|$N|O]]|$A|11|C|12|2|E|F|1F|G|@]|H|@]|6|$]]|$A|13|C|-4|2|E|F|1G|G|@]|H|@]|6|$]]]]

tf.placeholder() is meant to be fed to the session that when run receive the values from feed dict and perform the required operation.
Generally, you would create a Session() with 'with' keyword and run it. But this might not favour all situations due to which you would require immediate execution. This is called eager execution.
Example:
generally, this is the procedure to run a Session:
<pre><code>import tensorflow as tf

def square(num):
 return tf.square(num) 

p = tf.placeholder(tf.float32)
q = square(num)

with tf.Session() as sess:
 print(sess.run(q, feed_dict={num: 10})
</code></pre>
But when we run with eager execution we run it as:
<pre><code>import tensorflow as tf

tf.enable_eager_execution()

def square(num):
 return tf.square(num)

print(square(10)) 
</code></pre>
Therefore we need not run it inside a session explicitly and can be more intuitive in most of the cases. This provides more of an interactive execution.
For further details visit:
<a href="https://www.tensorflow.org/guide/eager" rel="noreferrer">https://www.tensorflow.org/guide/eager</a>
If you are converting the code from tensorflow v1 to tensorflow v2, You must implement tf.compat.v1 and Placeholder is present at tf.compat.v1.placeholder but this can only be executed in eager mode off.
<pre><code>tf.compat.v1.disable_eager_execution()
</code></pre>
TensorFlow released the eager execution mode, for which each node is immediately executed after definition. Statements using tf.placeholder are thus no longer valid.

entityMap|0|type|LINK|mutability|MUTABLE|data|url|https://www.tensorflow.org/guide/eager|1|https://www.tensorflow.org/beta/guide/effective_tf2|blocks|key|t6ne|text|在TensorFlow+1.X中，将创建占位符，并在实例化tf.Session时向其提供实际值。然而，从TensorFlow2.0开始，默认情况下已经启用了Eager+Execution，因此“占位符”的概念没有意义，因为操作是立即计算的(而不是与旧的范例不同)。|unstyled|depth|inlineStyleRanges|offset|length|style|CODE|entityRanges|95rb5|另请参阅Functions,+not+Sessions，|9auda|#+TensorFlow+1.X
outputs+=+session.run(f(placeholder),+feed_dict={placeholder:+input})
#+TensorFlow+2.0
outputs+=+f(input)|code-block|syntax|javascript|7ga0s^0|T|A|26|F|0|0|4|N|1|0|0^^$0|$1|$2|3|4|5|6|$7|8]]|9|$2|3|4|5|6|$7|A]]]|B|@$C|D|E|F|2|G|H|W|I|@$J|X|K|Y|L|M]]|N|@$J|Z|K|10|C|11]]|6|$]]|$C|O|E|P|2|G|H|12|I|@]|N|@$J|13|K|14|C|15]]|6|$]]|$C|Q|E|R|2|S|H|16|I|@]|N|@]|6|$T|U]]|$C|V|E|-4|2|G|H|17|I|@]|N|@]|6|$]]]]

In TensorFlow 1.X, placeholders are created and meant to be fed with actual values when a <code>tf.Session</code> is instantiated. However, from TensorFlow2.0 onwards, <a href="https://www.tensorflow.org/guide/eager" rel="noreferrer">Eager Execution</a> has been enabled by default, so the notion of a "placeholder" does not make sense as operations are computed immediately (rather than being differed with the old paradigm). 

Also see <a href="https://www.tensorflow.org/beta/guide/effective_tf2" rel="noreferrer">Functions, not Sessions</a>, 

<pre><code># TensorFlow 1.X
outputs = session.run(f(placeholder), feed_dict={placeholder: input})
# TensorFlow 2.0
outputs = f(input)
</code></pre>

entityMap|0|type|LINK|mutability|MUTABLE|data|url|https://stackoverflow.com/questions/53429896/disable-tensorflow-eager-execution|blocks|key|301tm|text|我在这里找到了一个简单的解决方案：disable+Tensorflow+eager+execution|unstyled|depth|inlineStyleRanges|entityRanges|offset|length|9k06k|基本上是这样的：|ator4|tf.compat.v1.disable_eager_execution()|style|CODE|dt06j|这样，您就禁用了默认的激活紧急执行，并且不需要更多地接触代码。|fbu09^0|H|Y|0|0|0|0|12|0|0^^$0|$1|$2|3|4|5|6|$7|8]]]|9|@$A|B|C|D|2|E|F|T|G|@]|H|@$I|U|J|V|A|W]]|6|$]]|$A|K|C|L|2|E|F|X|G|@]|H|@]|6|$]]|$A|M|C|N|2|E|F|Y|G|@$I|Z|J|10|O|P]]|H|@]|6|$]]|$A|Q|C|R|2|E|F|11|G|@]|H|@]|6|$]]|$A|S|C|-4|2|E|F|12|G|@]|H|@]|6|$]]]]

I found an easy solution here: <a href="https://stackoverflow.com/questions/53429896/disable-tensorflow-eager-execution">disable Tensorflow eager execution</a>

Basicaly it is:

<code>tf.compat.v1.disable_eager_execution()</code>

With this, you disable the default activate eager execution and you don't need to touch the code much more.

entityMap|0|type|LINK|mutability|MUTABLE|data|url|https://tensorflow-object-detection-api-tutorial.readthedocs.io/en/latest/training.html#exporting-a-trained-model|blocks|key|8pap8|text|如果您在使用TensorFlow模型进行对象检测时遇到此错误，则使用exporter_main_v2.py而不是export_inference_graph.py导出模型。这是正确的方法。如果您只是关闭eager_execution，那么它将解决此错误，但会生成其他错误。|unstyled|depth|inlineStyleRanges|offset|length|style|CODE|entityRanges|8odd5|还请注意，有一些参数更改，例如您将指定检查点目录的路径，而不是检查点的路径。有关如何使用TensorFlow+V2进行对象检测的信息，请参阅this文档|eik05^0|Y|J|1K|P|0|1Y|4|0|0^^$0|$1|$2|3|4|5|6|$7|8]]]|9|@$A|B|C|D|2|E|F|P|G|@$H|Q|I|R|J|K]|$H|S|I|T|J|K]]|L|@]|6|$]]|$A|M|C|N|2|E|F|U|G|@]|L|@$H|V|I|W|A|X]]|6|$]]|$A|O|C|-4|2|E|F|Y|G|@]|L|@]|6|$]]]]

If you are getting this error while doing object detection using TensorFlow model then use <code>exporter_main_v2.py</code> instead of <code>export_inference_graph.py</code> for exporting the model. This is right method to do. If you just off eager_execution then it will solve this error but generate other.
Also note that there are some parameter change like hear you will specify the path to directory of checkpoint instead of path to checkpoint. refer to <a href="https://tensorflow-object-detection-api-tutorial.readthedocs.io/en/latest/training.html#exporting-a-trained-model" rel="noreferrer">this</a> document for how to do object detection with TensorFlow V2

I have upgraded with tf_upgrade_v2 TF1 code to TF2. I'm a noob with both. I got the next error:

<pre><code>RuntimeError: tf.placeholder() is not compatible with eager execution.
</code></pre>

I have some <code>tf.compat.v1.placeholder()</code>.

<pre><code>self.temperature = tf.compat.v1.placeholder_with_default(1., shape=())
self.edges_labels = tf.compat.v1.placeholder(dtype=tf.int64, shape=(None, vertexes, vertexes))
self.nodes_labels = tf.compat.v1.placeholder(dtype=tf.int64, shape=(None, vertexes))
self.embeddings = tf.compat.v1.placeholder(dtype=tf.float32, shape=(None, embedding_dim))
</code></pre>

Could you give me any advice about how to proceed? Any "fast" solutions? or should I to recode this?

RuntimeError: tf.placeholder() is not compatible with eager execution

 我已经将tf_upgrade_v2 TF1代码升级到了TF2。我两个都是新手。我得到了下一个错误： RuntimeError: tf.placeholder() is not compatible with eager execution. 我有一些tf.compat.v1.placeholder()。 self.t...

问RuntimeError: tf.placeholder()与急切执行不兼容
EN

社区

活动

资源

关于

腾讯云开发者

热门产品

热门推荐

更多推荐

问RuntimeError: tf.placeholder()与急切执行不兼容EN

社区

活动

资源

关于

腾讯云开发者

热门产品

热门推荐

更多推荐

问RuntimeError: tf.placeholder()与急切执行不兼容
EN