使用keras时出现内存错误

Question

我正在为CNN使用keras，但问题是存在内存泄漏。错误是

        anushreej@cpusrv-gpu-109:~/12EC35005/MTP_Workspace/MTP$ python cnn_implement.py
        Using Theano backend.
        [INFO] compiling model...
        Traceback (most recent call last):
          File "cnn_implement.py", line 23, in <module>
            model = CNNModel.build(width=150, height=150, depth=3)
          File "/home/ms/anushreej/12EC35005/MTP_Workspace/MTP/cnn/networks/model_define.py", line 27, in build
            model.add(Dense(depth*height*width))
          File "/home/ms/anushreej/anaconda3/lib/python3.5/site-packages/keras/models.py", line 146, in add
            output_tensor = layer(self.outputs[0])
          File "/home/ms/anushreej/anaconda3/lib/python3.5/site-packages/keras/engine/topology.py", line 458, in __call__
            self.build(input_shapes[0])
          File "/home/ms/anushreej/anaconda3/lib/python3.5/site-packages/keras/layers/core.py", line 604, in build
            name='{}_W'.format(self.name))
          File "/home/ms/anushreej/anaconda3/lib/python3.5/site-packages/keras/initializations.py", line 61, in glorot_uniform
            return uniform(shape, s, name=name)
          File "/home/ms/anushreej/anaconda3/lib/python3.5/site-packages/keras/initializations.py", line 32, in uniform
            return K.variable(np.random.uniform(low=-scale, high=scale, size=shape),
          File "mtrand.pyx", line 1255, in mtrand.RandomState.uniform (numpy/random/mtrand/mtrand.c:13575)
          File "mtrand.pyx", line 220, in mtrand.cont2_array_sc (numpy/random/mtrand/mtrand.c:2902)
        MemoryError

现在我无法理解为什么会发生这种情况。我的训练图像非常小，大小为150*150*3。

代码是-:

        # import the necessary packages
        from keras.models import Sequential
        from keras.layers.convolutional import Convolution2D
        from keras.layers.core import Activation
        from keras.layers.core import Flatten
        from keras.layers.core import Dense
        class CNNModel:
          @staticmethod
          def build(width, height, depth):
            # initialize the model
            model = Sequential()
            # first set of CONV => RELU
            model.add(Convolution2D(50, 5, 5, border_mode="same", batch_input_shape=(None, depth, height, width)))
            model.add(Activation("relu"))
            # second set of CONV => RELU
            # model.add(Convolution2D(50, 5, 5, border_mode="same"))
            # model.add(Activation("relu"))
            # third set of CONV => RELU
            # model.add(Convolution2D(50, 5, 5, border_mode="same"))
            # model.add(Activation("relu"))
            model.add(Flatten())
            model.add(Dense(depth*height*width))
            # if weightsPath is not None:
            #   model.load_weights(weightsPath) 
            return model

Answer 1

我面临着同样的问题，我认为问题是在扁平化层之前的数据点数量超过了你的系统可以处理的数量(我在不同的系统中尝试过，所以一个具有高ram的系统工作，而ram较少的系统给出了这个错误)。只要添加更多的CNN层来减小尺寸，然后添加一个平坦层就可以了。

这给了我一个错误:

model = Sequential()
model.add(Convolution2D(32, 3, 3,border_mode='same',input_shape=(1, 96, 96),activation='relu'))
model.add(Convolution2D(64, 3, 3,border_mode='same',activation='relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(Flatten())
model.add(Dense(1000,activation='relu'))
model.add(Dense(97,activation='softmax'))

这没有给出一个错误

model = Sequential()
model.add(Convolution2D(32, 3, 3,border_mode='same',input_shape=(1, 96, 96),activation='relu'))
model.add(Convolution2D(64, 3, 3,border_mode='same',activation='relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(Convolution2D(64, 3, 3,border_mode='same',activation='relu'))
model.add(Convolution2D(128, 3, 3,border_mode='same',activation='relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(Flatten())
model.add(Dense(1000,activation='relu'))
model.add(Dense(97,activation='softmax')

希望能有所帮助。