TensorFlow——在web.py、Django环境下TensorFlow（Keras、tf.keras）加载和使用多模型失败解决方案

逃离我推掉我的手 2023-01-11 13:08 80阅读 0赞

# 问题描述 #

> Cannot interpret feed\_dict key as Tensor: Tensor Tensor("Placeholder\_8:0", shape=(3, 3, 128, 256), dtype=float32) is not an element of this graph

> ValueError: Variable conv1/weights already exists, disallowed. Did you mean to set reuse=True in VarScope? Originally defined at:

> Tensor Tensor("Placeholder:0", shape=(3, 3, 3, 32), dtype=float32) is not an element of this graph.

![SouthEast][]

# 问题分析 #

在Tensorflow中，所有操作对象都包装到相应的Session中的，所以想要使用不同的模型就需要将这些模型加载到不同的Session中并在使用的时候申明是哪个Session，从而避免由于Session和想使用的模型不匹配导致的错误。

而使用多个graph，就需要为每个graph使用不同的Session，但是每个graph也可以在多个Session中使用，这个时候就需要在每个Session使用的时候明确申明使用的graph。

# 解决方案 #

方法一：

当建立一个**from keras.models import Sequential的序贯模型或者是from keras.models import Model的函数式模型的model时**

import keras
    keras.backend.clear_session()

![20190817151552262.png][]

当然这似乎不是一种正常的解决思路。

方法二：

加载模型后，先执行一次model.predict()操作，之后的调用就不会出问题了

model = load_model(filepath=model_path)
    y = model.predict(x)

> 因此，在大多数的建立模型立即predict的demo测试中，并不出现问题。

这似乎也不是一种正常的解决思路。

方法三：

[TensorFlow——加载和使用多个模型解决方案][TensorFlow]

g1 = tf.Graph() # 加载到Session 1的graph
    g2 = tf.Graph() # 加载到Session 2的graph
     
    sess1 = tf.Session(graph=g1) # Session1
    sess2 = tf.Session(graph=g2) # Session2
     
    # 加载第一个模型
    with sess1.as_default(): 
        with g1.as_default():
            tf.global_variables_initializer().run()
            model_saver = tf.train.Saver(tf.global_variables())
            model_ckpt = tf.train.get_checkpoint_state(“model1/save/path”)
            model_saver.restore(sess, model_ckpt.model_checkpoint_path)
    # 加载第二个模型
    with sess2.as_default():  # 1
        with g2.as_default():  
            tf.global_variables_initializer().run()
            model_saver = tf.train.Saver(tf.global_variables())
            model_ckpt = tf.train.get_checkpoint_state(“model2/save/path”)
            model_saver.restore(sess, model_ckpt.model_checkpoint_path)
     
    ...
     
    # 使用的时候
    with sess1.as_default():
        with sess1.graph.as_default():  # 2
            ...
     
    with sess2.as_default():
        with sess2.graph.as_default():
            ...
     
    # 关闭sess
    sess1.close()
    sess2.close()

# 参考文章 #

[TensorFlow——加载和使用多个模型解决方案][TensorFlow]

[https://stackoverflow.com/questions/41607144/loading-two-models-from-saver-in-the-same-tensorflow-session][https_stackoverflow.com_questions_41607144_loading-two-models-from-saver-in-the-same-tensorflow-session]

[https://stackoverflow.com/questions/39175945/run-multiple-pre-trained-tensorflow-nets-at-the-same-time][https_stackoverflow.com_questions_39175945_run-multiple-pre-trained-tensorflow-nets-at-the-same-time]

[Cannot interpret feed\_dict key as Tensor: Tensor Tensor("Placeholder\_8:0", shape=(3, 3, 128, 256), d][Cannot interpret feed_dict key as Tensor_ Tensor Tensor_Placeholder_8_0_ shape_3_ 3_ 128_ 256_ d]

[ValueError: Tensor Tensor("Placeholder:0", shape=(3, 3, 1, 32), dtype=float32)][ValueError_ Tensor Tensor_Placeholder_0_ shape_3_ 3_ 1_ 32_ dtype_float32]

[解决在django中应用keras模型时出现的ValueError("Tensor %s is not an element of this graph." % obj)问题][django_keras_ValueError_Tensor _s is not an element of this graph._ _ obj]

[将keras模型在django中应用时出现的小问题][keras_django]

[Django使用Keras的几个坑][Django_Keras]

[解决Django中调用keras的模型出现的问题][Django_keras]

https://blog.csdn.net/weixin\_30034903/article/details/106144194

https://blog.csdn.net/xxzhix/article/details/81983982

[\[2020/05/12 更新\] Django调用Keras的模型][2020_05_12 _ Django_Keras]

[tensorflow多次restore模型出错][tensorflow_restore]

[Keras与Django整合问题（keras反复识别问题）][Keras_Django_keras]

[https://www.freesion.com/article/7982900117/][https_www.freesion.com_article_7982900117]

[SouthEast]: /images/20221119/53e6f55ddef842f6a5126074a5c30492.png
[20190817151552262.png]: /images/20221119/9ae2fbf306094438803ad42cb1a31d75.png
[TensorFlow]: https://shentuzhigang.blog.csdn.net/article/details/113206252
[https_stackoverflow.com_questions_41607144_loading-two-models-from-saver-in-the-same-tensorflow-session]: https://stackoverflow.com/questions/41607144/loading-two-models-from-saver-in-the-same-tensorflow-session
[https_stackoverflow.com_questions_39175945_run-multiple-pre-trained-tensorflow-nets-at-the-same-time]: https://stackoverflow.com/questions/39175945/run-multiple-pre-trained-tensorflow-nets-at-the-same-time
[Cannot interpret feed_dict key as Tensor_ Tensor Tensor_Placeholder_8_0_ shape_3_ 3_ 128_ 256_ d]: https://blog.csdn.net/selfimpro_001/article/details/99693273
[ValueError_ Tensor Tensor_Placeholder_0_ shape_3_ 3_ 1_ 32_ dtype_float32]: https://blog.csdn.net/qq_37879432/article/details/79032664
[django_keras_ValueError_Tensor _s is not an element of this graph._ _ obj]: https://www.cnblogs.com/bymo/p/8955009.html
[keras_django]: https://zhuanlan.zhihu.com/p/27101000
[Django_Keras]: https://blog.csdn.net/weixin_30034903/article/details/106144194
[Django_keras]: https://www.jb51.net/article/167151.htm
[2020_05_12 _ Django_Keras]: https://blog.csdn.net/sinat_41657218/article/details/103730727
[tensorflow_restore]: https://blog.csdn.net/xxzhix/article/details/81983982
[Keras_Django_keras]: https://blog.csdn.net/qq_38003892/article/details/79073821
[https_www.freesion.com_article_7982900117]: https://www.freesion.com/article/7982900117/