Django整合Keras報錯：ValueError: Tensor Tensor("Placeholder:0", shape=(3, 3, 1, 32), dtype=float32) is no

本人在寫Django RESful API時，碰到一個難題，老出現，整合Keras,報以下錯誤；很糾結，探索找資料近一個星期，皇天不負有心人，解決了

 

Internal Server Error: /pic/analysis/
Traceback (most recent call last):
  File "D:\AI\Python35\lib\site-packages\tensorflow\python\client\session.py", line 1064, in _run
    allow_operation=False)
  File "D:\AI\Python35\lib\site-packages\tensorflow\python\framework\ops.py", line 3035, in as_graph_element
    return self._as_graph_element_locked(obj, allow_tensor, allow_operation)
  File "D:\AI\Python35\lib\site-packages\tensorflow\python\framework\ops.py", line 3114, in _as_graph_element_locked
    raise ValueError("Tensor %s is not an element of this graph." % obj)
ValueError: Tensor Tensor("Placeholder:0", shape=(3, 3, 1, 32), dtype=float32) is not an element of this graph.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "D:\AI\Python35\lib\site-packages\django\core\handlers\exception.py", line 35, in inner
    response = get_response(request)
  File "D:\AI\Python35\lib\site-packages\django\core\handlers\base.py", line 128, in _get_response
    response = self.process_exception_by_middleware(e, request)
  File "D:\AI\Python35\lib\site-packages\django\core\handlers\base.py", line 126, in _get_response
    response = wrapped_callback(request, *callback_args, **callback_kwargs)
  File "D:\AI\Python35\lib\site-packages\django\test\utils.py", line 371, in inner
    return func(*args, **kwargs)
  File "E:\PyCharmWork\PythonWebApp\ApiPicDjangoSite\PicWeb\views.py", line 25, in picAnalysis
    predictBank(im, name)  #對圖片進行預測，並返回預測結果
  File "E:\PyCharmWork\PythonWebApp\ApiPicDjangoSite\PicWeb\PreDictBankFunc.py", line 42, in predictBank
    model = load_model('static\CnnBankUp.h5', compile=False)
  File "D:\AI\Python35\lib\site-packages\keras\models.py", line 243, in load_model
    topology.load_weights_from_hdf5_group(f['model_weights'], model.layers)
  File "D:\AI\Python35\lib\site-packages\keras\engine\topology.py", line 3142, in load_weights_from_hdf5_group
    K.batch_set_value(weight_value_tuples)
  File "D:\AI\Python35\lib\site-packages\keras\backend\tensorflow_backend.py", line 2247, in batch_set_value
    get_session().run(assign_ops, feed_dict=feed_dict)
  File "D:\AI\Python35\lib\site-packages\tensorflow\python\client\session.py", line 889, in run
    run_metadata_ptr)
  File "D:\AI\Python35\lib\site-packages\tensorflow\python\client\session.py", line 1067, in _run
    + e.args[0])
TypeError: Cannot interpret feed_dict key as Tensor: Tensor Tensor("Placeholder:0", shape=(3, 3, 1, 32), dtype=float32) is not an element of t
his graph.
[08/Jan/2018 08:53:15] "POST /pic/analysis/ HTTP/1.1" 500 159200

 

我的理解：在初始化Django時，把keras中 model先初始化，省得後面不斷調用，產生莫名其妙的問題

Django初始化代碼寫在__init__.py中：

from keras.models import load_model
import numpy as np

#參考：https://zhuanlan.zhihu.com/p/27101000

print('load model...')
model = load_model('static\\CnnBankUp.h5', compile=False)
print('load done.')

#必定要添加這段代碼，先測試一下，能夠避免ValueError: Tensor Tensor("Placeholder:0", shape=(3, 3, 1, 32), dtype=float32) 
#is not an element of this graph.的錯誤
print('test model...')
#根據本身傳入圖片格式定義np.zeros（）
print(model.predict(np.zeros((2, 200,200,1))))
print('test done.')


# 使用模型，在獲得用戶輸入時會調用如下兩個函數進行實時文本分類
# 輸入參數 comment 爲通過了分詞與向量化處理後的模型輸入
def picType_class(comment):
    global model
    result_vec=None
    result_vec = model.predict(comment)
    return result_vec

 

頗有用的參考，很是感謝

參考：https://zhuanlan.zhihu.com/p/27101000