Does the input size of sequenceInputLayer have to be set 1?

4 次查看（过去 30 天）
Jack Xiao
Jack Xiao 2022-4-11
回答： Zhiyu WANG 2022-5-25
i followed the example https://ww2.mathworks.cn/help/textanalytics/ug/classify-text-data-using-deep-learning.html.
The network is as follows:
inputSize = 1;
embeddingDimension = 50;
numHiddenUnits = 80;
numWords = enc.NumWords;
numClasses = numel(categories(YTrain));
layers = [ ...
sequenceInputLayer(inputSize)
wordEmbeddingLayer(embeddingDimension,numWords)
lstmLayer(numHiddenUnits,'OutputMode','last')
fullyConnectedLayer(numClasses)
softmaxLayer
classificationLayer]
I only tried to change inputSize = 2, then the network can not be trained with the error:
Does the input size of sequenceInputLayer have to be set 1?

请先登录，再回答此问题。

回答（1 个）

Zhiyu WANG
Zhiyu WANG 2022-5-25
Becasue the input sequence data is a 1 by 10 double.It menas each time you take 1 token out of the sequence.

类别

在 Help Center 和 File Exchange 中查找有关 Deep Learning Toolbox 的更多信息

标签

产品


版本

R2021b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by