在图表上调用finalize()可以帮助您.

def __init__(self,model_path):
        self.cnn_model = load_model(model_path)
        self.session = K.get_session()
        self.graph = tf.get_default_graph()
        self.graph.finalize()

更新1：finalize()将使您的图形为只读,以便可以安全地在多个线程中使用.作为副作用,它将帮助您找到无意的行为,有时还会发现内存泄漏,因为当您尝试修改图形时它会引发异常.

想象一下,你有一个线程可以做一个例如输入的热编码. (坏的例子:)

def preprocessing(self,data):
    one_hot_data = tf.one_hot(data,depth=self.num_classes)
    return self.session.run(one_hot_data)

如果在图表中打印对象数量,您会发现它会随着时间的推移而增加

# amount of nodes in tf graph
print(len(list(tf.get_default_graph().as_graph_def().node)))

但是,如果您首先定义图形不是这种情况(略微更好的代码)：

def preprocessing(self,data):
    # run pre-created operation with self.input as placeholder
    return self.session.run(self.one_hot_data,Feed_dict={self.input: data})

更新2：根据此thread,您需要在执行多线程之前在keras模型上调用model._make_predict_function().

Keras builds the GPU function the first time you call predict(). That
way,if you never call predict,you save some time and resources.
However,the first time you call predict is slightly slower than every
other time.

更新的代码：

def __init__(self,model_path):
    self.cnn_model = load_model(model_path)
    self.cnn_model._make_predict_function() # have to initialize before threading
    self.session = K.get_session()
    self.graph = tf.get_default_graph() 
    self.graph.finalize() # make graph read-only

更新3：我做了一个预热概念的证明,因为_make_predict_function()似乎没有按预期工作.
首先我创建了一个虚拟模型：

import tensorflow as tf
from keras.layers import *
from keras.models import *

model = Sequential()
model.add(Dense(256,input_shape=(2,)))
model.add(Dense(1,activation='softmax'))

model.compile(loss='mean_squared_error',optimizer='adam')

model.save("dummymodel")

然后在另一个脚本中我加载了该模型并使其在多个线程上运行

import tensorflow as tf
from keras import backend as K
from keras.models import load_model
import threading as t
import numpy as np

K.clear_session()

class CNN:
    def __init__(self,model_path):

        self.cnn_model = load_model(model_path)
        self.cnn_model.predict(np.array([[0,0]])) # warmup
        self.session = K.get_session()
        self.graph = tf.get_default_graph()
        self.graph.finalize() # finalize

    def preproccesing(self,data):
        # dummy
        return data

    def query_cnn(self,data):
        X = self.preproccesing(data)
        with self.session.as_default():
            with self.graph.as_default():
                prediction = self.cnn_model.predict(X)
        print(prediction)
        return prediction


cnn = CNN("dummymodel")

th = t.Thread(target=cnn.query_cnn,kwargs={"data": np.random.random((500,2))})
th2 = t.Thread(target=cnn.query_cnn,2))})
th3 = t.Thread(target=cnn.query_cnn,2))})
th4 = t.Thread(target=cnn.query_cnn,2))})
th5 = t.Thread(target=cnn.query_cnn,2))})
th.start()
th2.start()
th3.start()
th4.start()
th5.start()

th2.join()
th.join()
th3.join()
th5.join()
th4.join()

评论预热和最终确定的线条我能够重现你的第一个问题