为什么keras不允许以这种方式添加卷积层? [英] Why keras does not allow to add a convolutional layer in this way?

问题描述

以下代码

from tensorflow import keras
from keras.layers import Conv2D

model = keras.Sequential()
model.add(Conv2D(1, (3, 3), padding='same', input_shape=(28, 28, 1)))

执行时抛出错误:

TypeError: The added layer must be an instance of class Layer. Found: <keras.layers.convolutional.Conv2D object at 0x7fea0c002d10>

我也尝试使用Convolutional2D，但出现了相同的错误.为什么?

I also tried using the Convolutional2D but got the same error. Why?

推荐答案

尝试一下:

from tensorflow import keras
from tensorflow.keras.layers import Conv2D

model = keras.Sequential()
model.add(Conv2D(1, (3, 3), padding='same', input_shape=(28, 28, 1)))

您正在将tf.keras顺序模型与keras Conv2D层(而不是tf.keras Conv2D层)混合在一起.

You are mixing a tf.keras Sequential model with a keras Conv2D layer (instead of a tf.keras Conv2D layer.)

或者，如下所述，使用实际的Keras:

Or, as remarked below, use actual Keras:

import keras
from keras.models import Sequential
from keras.layers import Conv2D

model = Sequential()
model.add(Conv2D(1, (3, 3), padding='same', input_shape=(28, 28, 1)))

这篇关于为什么keras不允许以这种方式添加卷积层?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！