keras中的Conv2D层输出形状答案

【问题标题】：Conv2D layer output shape in keraskeras中的Conv2D层输出形状
【发布时间】：2017-09-12 11:03:07
【问题描述】：

我想手动输入 keras Conv2D 层。

我采用 MNIST 数据集。

Conv2D 只接受张量，所以我使用 keras 的 Input 命令将 x_train 更改为 x_train_tensor。

我的输入是 keras 指令中给出的格式

(samples,rows, cols,channels)

示例输入：

(60000,128,128,1)

我希望输出类似于：

(None, 26, 26, 32)

我得到：

shape=(?, 59998, 26, 32)

我做错了什么？

我的代码：

import keras
from keras.datasets import mnist
from keras.layers import Conv2D
from keras import backend as K
from keras.layers import Input
batch_size = 128
num_classes = 10
epochs = 1
# input image dimensions
img_rows, img_cols = 28, 28
# the data, shuffled and split between train and test sets
(x_train, y_train), (x_test, y_test) = mnist.load_data()
if K.image_data_format() == 'channels_first':
    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)
    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)
    input_shape = (1, img_rows, img_cols)
else:
    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)
    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)
    input_shape = (img_rows, img_cols, 1)

x_train = x_train.astype('float32')
x_test = x_test.astype('float32')
x_train /= 255
x_test /= 255
print('x_train shape:', x_train.shape)
print(x_train.shape[0], 'train samples')
print(x_test.shape[0], 'test samples')

x_train_tensor=Input(shape=(60000,28,28), name='x_train')
A=Conv2D(32, kernel_size=(3, 3),
       activation='relu',
       input_shape=input_shape)(x_train_tensor)

【问题讨论】：

标签： python-3.x keras

【解决方案1】：

样本数量不是input_shape 的一部分，在您的情况下，您犯了两个错误。首先是输入形状错误，其次是指定两个输入形状，一次在 Input 构造函数中，第二次在 Conv2D 实例中：

x_train_tensor=Input(shape=(28, 28, 1), name='x_train')
A=Conv2D(32, kernel_size=(3, 3), activation='relu')(x_train_tensor)

【讨论】：