无法添加第四个卷积层

Question

我对机械学习很陌生，当我在看卷积神经网络的教程时，我想自己实验如何提高精度。然而，当我试图在我的模型中添加另一个卷积和池化层时，它显示了一条错误消息。这是在我添加层之前：

model = models.Sequential()
model.add(layers.Conv2D(32, (3, 3), activation='relu', input_shape=(28, 28, 1)))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.Flatten())
model.add(layers.Dense(64, activation='relu'))
model.add(layers.Dense(62))

这是在之后

model = models.Sequential()
model.add(layers.Conv2D(32, (3, 3), activation='relu', input_shape=(28, 28, 1)))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.Flatten())
model.add(layers.Dense(64, activation='relu'))
model.add(layers.Dense(62))

这是它给我的错误信息：

ValueError：对于"｛｛node conv2d_36/conv2d｝｝=conv2d[T=DT_FLOAT，data_format="NHWC"，expansions=[1，1，1，1]，explicit_paddings=[]，paddings="；"VALID"；，progress=[1,1,11]，use_cudnn_on_gpu=true](max_pooling2d_26/MaxPool，conv2d_36/conv2d/ReadVariableOp('，输入形状：[？，1,1,64]，[3,3,64,64]。网站：stackoverflow.com

Answer 1

这是因为你在网络内部减少了太多维度。在卷积层中使用padding='same'来避免这种维度错误

model = models.Sequential()
model.add(layers.Conv2D(32, (3, 3), activation='relu', padding='same', 
input_shape=(28, 28, 1)))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu', padding='same'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu', padding='same'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu', padding='same'))
model.add(layers.Flatten())
model.add(layers.Dense(64, activation='relu'))
model.add(layers.Dense(62))
model.summary()