Feeds
質問
is it possible to make "Sequence-to-Sequence Classification" using custom CNN network?
If so , how to create a sequence datastore and what`s the dataformat of the corresponding dlarray object?
3年以上 前 | 1 件の回答 | 0
1
回答回答済み
SequenceinputLayer and Convolution2dLayer
It`s recommended to transform your data to formatted dlarray objects at first. As for the dlarray object, you can refer to this...
SequenceinputLayer and Convolution2dLayer
It`s recommended to transform your data to formatted dlarray objects at first. As for the dlarray object, you can refer to this...
3年以上 前 | 0
回答済み
using predictAndUpdateState with an LSTM with featureInputLayer
LSTM requires somelike "timeseries" data while "featureInputLayer" is not along with that.Yo ucan refer to this answer: https:/...
using predictAndUpdateState with an LSTM with featureInputLayer
LSTM requires somelike "timeseries" data while "featureInputLayer" is not along with that.Yo ucan refer to this answer: https:/...
3年以上 前 | 0
| 採用済み
回答済み
Does the input size of sequenceInputLayer have to be set 1?
Becasue the input sequence data is a 1 by 10 double.It menas each time you take 1 token out of the sequence.
Does the input size of sequenceInputLayer have to be set 1?
Becasue the input sequence data is a 1 by 10 double.It menas each time you take 1 token out of the sequence.
3年以上 前 | 0
