LSTM Keras中的尺寸不匹配 [英] Dimension Mismatch in LSTM Keras

问题描述

我想创建一个可以添加两个字节的基本RNN.这是输入和输出，只需简单的加法即可

I want to create a basic RNN that can add two bytes. Here are the input and outputs, which are expected of a simple addition

X = [[0, 0], [0, 1], [1, 1], [0, 1], [1, 0], [1, 0], [1, 1], [1, 0]]

即X1 = 00101111和X2 = 01110010

Y = [1, 0, 1, 0, 0, 0, 0, 1]

我创建了以下顺序模型

model = Sequential()
model.add(GRU(output_dim = 16, input_length = 2, input_dim = 8))
model.add(Activation('relu'`))
model.add(Dense(2, activation='softmax'))
model.compile(loss = 'binary_crossentropy', optimizer = 'adam', metrics = ['accuracy'])
model.summary()

我遇到的错误是

预计lstm_input_1具有3个维度，但数组的形状为(8L, 2L)

expected lstm_input_1 to have 3 dimensions, but got array with shape (8L, 2L)

因此，如果我通过将X更改为

So if I increase the dimensions by changing X to

[[[0 0]] [[1 1]] [[1 1]] [[1 0]] [[0 0]] [[1 0]] [[0 1]] [[1 0]]]

然后错误更改为

期望lstm_input_1具有形状(None, 8, 2)，但具有形状为(8L, 1L, 2L)

expected lstm_input_1 to have shape (None, 8, 2) but got array with shape (8L, 1L, 2L)

推荐答案

将X更改为[[[0, 0], [0, 1], [1, 1], [0, 1], [1, 0], [1, 0], [1, 1], [1, 0]]]，使其形状为(1, 8, 2)

Change X to [[[0, 0], [0, 1], [1, 1], [0, 1], [1, 0], [1, 0], [1, 1], [1, 0]]] so that its shape is (1, 8, 2)

这篇关于LSTM Keras中的尺寸不匹配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！