Tensorflow 卷积神经网络负维大小

Question

I am building this CNN model, '''

x_test = np.array(x_test).reshape(-1,IMAGE_SIZE,IMAGE_SIZE,1)
x_train = np.array(x_train).reshape(-1,IMAGE_SIZE,IMAGE_SIZE,1)


model = tf.keras.Sequential()

model.add(Conv2D(16,(3,3),padding='same',input_shape=x_train.shape[1:]))  
model.add(Activation("relu"))
model.add(MaxPooling2D(pool_size=(2,2)))

model.add(Conv2D(32,(3,3),strides=(2,2),padding='valid'))
model.add(Activation("relu"))
model.add(MaxPooling2D(pool_size=(2,2)))

model.add(Conv2D(32,(2,2),padding='same')) 
model.add(Activation("relu"))
model.add(MaxPooling2D(pool_size=(2,2)))

model.add(Conv2D(64,(2,2),strides=(2,2),padding='valid'))
model.add(Activation("relu"))

model.add(GlobalAveragePooling2D())
model.add(Dense(64))
model.add(Activation("relu"))
model.add(Dense(10))
model.add(Activation("softmax"))

'''

But this is giving me an error:- InvalidArgumentError: Negative dimension size caused by subtracting 2 from 1 for '{{node conv2d_115/Conv2D}} = Conv2D[T=DT_FLOAT, data_format="NHWC", dilations=[1, 1, 1, 1], explicit_paddings=[], padding="VALID", strides=[1, 2, 2, 1], use_cudnn_on_gpu=true](Placeholder, conv2d_115/Conv2D/ReadVariableOp)' with input shapes: [?,1,1,32], [2,2,32,64].

During handling of the above exception, another exception occurred:

ValueError: Negative dimension size caused by subtracting 2 from 1 for '{{node conv2d_115/Conv2D}} = Conv2D[T=DT_FLOAT, data_format="NHWC", dilations=[1, 1, 1, 1], explicit_paddings=[], padding="VALID", strides=[1, 2, 2, 1], use_cudnn_on_gpu=true](Placeholder, conv2d_115/Conv2D/ReadVariableOp)' with input shapes: [?,1,1,32], [2,2,32,64].

What does this error mean? What I am doing wrong? Please help me in fixing this.

Answer 1

这可能意味着，由于像 maxpooling 和 conv2d 这样的层具有“有效”填充值，某些层的输入变得非常小，以至于张量的形状为 [?,1,1,32]，因此它不能执行大小为 2 的 2D maxpooling 或 2D 卷积。在这种情况下，您需要根据输入大小重新设计层（意味着更改层数，或输入大小、步幅、填充选项等） .