Training a multi-layer neural network for (tabular) BigData

1 回表示 (過去 30 日間)

Ramakrishnan Raman 2018 年 9 月 3 日

0
リンク

この質問への直接リンク

https://jp.mathworks.com/matlabcentral/answers/417220-training-a-multi-layer-neural-network-for-tabular-bigdata

コメント済み: Giorgio De Nunzio 2020 年 4 月 25 日

How to train a multi-layer neural network (for instance, using feedforwardnet) using BigData (for instance, using TallArrays). Shouldn't it be supported to pass the tall array object containing features/target to the training function and have the training happen? The data input format is a simple tabular(matrix data) - i.e. not image or sequence or time series data

1 件のコメント
-1 件の古いコメントを表示-1 件の古いコメントを非表示

Giorgio De Nunzio 2020 年 4 月 25 日

Hi, did you find how to work with neural networks and tall arrays? I am tryng now. Moreover, patternnet train wants features on the rows, not samples, and it is not possible to transpose tall arrays (mine are classical one-line-per-sample tables with each column giving a feature: I can transpose normal matrices, but not tall arrays). Thanks. Giorgio

サインインしてコメントする。

サインインしてこの質問に回答する。