在Flask框架下解决Tensorlfow模型加载无效问题

在Flask框架下，加载tensrflow模型会出现如下错误

"The name 'encoder_inputs' refers to an Operation not in the graph."

原因在于tf模型与flask并不在同一个线程中。解决方案参考tf官网给出的：

 

tf.get_default_graph

Defined in tensorflow/python/framework/ops.py.

Returns the default graph for the current thread.

The returned graph will be the innermost graph on which a Graph.as_default() context has been entered, or a global default graph if none has been explicitly created.

NOTE: The default graph is a property of the current thread. If you create a new thread, and wish to use the default graph in that thread, you must explicitly add a with g.as_default(): in that thread's function.

利用NOTE给的方法，使用当前线程

graph = tf.get_default_graph()

with graph.as_default():# 这一行代码一定要
    xxx
    xxx

即可解决问题