python - 输入形状为 [?,1,2,36], [5,5,36,48] 的“conv2d_7/Conv2D”（操作：“Conv2D”）从 1 中减去 5 导致的负尺寸大小

Question

我已经使用来自 Tensorflow 的 Keras（版本 = 2.2.4-tf）构建了以下模型

model = tf.keras.Sequential()
model.add(Conv2D(24, 5, 5, padding='same', kernel_initializer='he_normal', input_shape = (150,200, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D())
model.add(Activation('relu'))
model.add(GlobalAveragePooling2D())
model.add(Dense(18))

但不知何故，我收到以下错误：Negative dimension size caused by subtracting 5 from 1 for 'conv2d_7/Conv2D' (op: 'Conv2D') with input shapes: [?,1,2,36], [5,5,36,48].

标签： pythontensorflowmachine-learningkeras