How to configure input dataset in RNN/LSTM Many to One problem

Hello everyone

I’m currently working on RNN/LSTM many to one classification problem.

There are a 4 labels in total for classfication, and each label consists of a dataset with sizes of [30(time step), 6,500~11,000(batch size), 4(input(feature) size)].

In order to designate one label per matrix of [30 x 4] to organize the dataset required for training (x, y(label)), how should I configure the data set in the pytorch environment?