运行Keras和scikitlearn时发生了不同的异常

Question

我尝试将Keras模型（作为函数）传递给scikit\u learn的KerasClassifier包装器，然后使用GridSearchCV创建一些设置，最后拟合train和test数据集（都是numpy数组）

然后，使用相同的python脚本，我得到了不同的异常，其中一些异常是：

第1部分。你知道吗

Traceback (most recent call last): File "mnist_flat_imac.py", line 63, in grid_result = validator.fit(train_images, train_labels) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/site-packages/sklearn/model_selection/_search.py", line 626, in fit base_estimator = clone(self.estimator) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/site-packages/sklearn/base.py", line 62, in clone new_object_params[name] = clone(param, safe=False) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/site-packages/sklearn/base.py", line 53, in clone

snipped here

in _deepcopy_dict y[deepcopy(key, memo)] = deepcopy(value, memo) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/copy.py", line 174, in deepcopy rv = reductor(4) TypeError: can't pickle SwigPyObject objects Exception ignored in: > Traceback (most recent call last): File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/site-packages/tensorflow/python/framework/c_api_util.py", line 52, in __del__ c_api.TF_DeleteGraph(self.graph) AttributeError: 'ScopedTFGraph' object has no attribute 'graph'

第2部分。你知道吗

Traceback (most recent call last): File "mnist_flat_imac.py", line 63, in grid_result = validator.fit(train_images, train_labels) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/site-packages/sklearn/model_selection/_search.py", line 626, in fit base_estimator = clone(self.estimator) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/site-packages/sklearn/base.py", line 62, in clone new_object_params[name] = clone(param, safe=False) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/site-packages/sklearn/base.py", line 53, in clone return copy.deepcopy(estimator) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/copy.py", line 182, in deepcopy y = _reconstruct(x, rv, 1, memo) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/copy.py", line 297, in _reconstruct

snipped here

in deepcopy y = _reconstruct(x, rv, 1, memo) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/copy.py", line 297, in _reconstruct state = deepcopy(state, memo) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/copy.py", line 155, in deepcopy y = copier(x, memo) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/copy.py", line 243, in _deepcopy_dict y[deepcopy(key, memo)] = deepcopy(value, memo) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/copy.py", line 174, in deepcopy rv = reductor(4) TypeError: can't pickle SwigPyObject objects

第三章。你知道吗

Traceback (most recent call last): File "mnist_flat_imac.py", line 63, in grid_result = validator.fit(train_images, train_labels) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/site-packages/sklearn/model_selection/_search.py", line 626, in fit base_estimator = clone(self.estimator) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/site-packages/sklearn/base.py", line 62, in clone new_object_params[name] = clone(param, safe=False) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/site-packages/sklearn/base.py", line 53, in clone

snipped here

in _deepcopy_dict y[deepcopy(key, memo)] = deepcopy(value, memo) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/copy.py", line 182, in deepcopy y = _reconstruct(x, rv, 1, memo) File "/home/longnv/PYTHON_ENV/DataScience/lib/python3.5/copy.py", line 306, in _reconstruct y.dict.update(state) AttributeError: 'NoneType' object has no attribute 'update'

为什么它用同一个python脚本输出不同的错误？ 我该怎么解决这个问题？你知道吗

非常感谢！你知道吗

附言

• python:3.5版
• 张量流：1.10.1
• 熊猫：0.23.4
• Ubuntu:4.4.0-124-generic

Answer 1

找到了。你知道吗

应该是： clf=KerasClassifier（构建\ fn=获取\模型）

而不是： clf=KerasClassifier（build\u fn=get\u model（））