How do you train nn.LSTM with multidimensional sequence data?

If i wanted to train nn.LSTM on sequence data, say from a csv that had multiple columns, like this:

timestamp columna columnb columnc columnd
10000, 1, 2, 3, 4
10001, 1, 2.1, 3, 4

100## 1, 2, 3, 4, label

where the sequence of values in the columns represents the change over time lets say 5 seconds, and the whole sequence represents the behavior i wish to learn and make predictions on, how would i do that? most of the examples i see for LSTM all show a single value’s changes over time. How does one present multiple dimensions to the LSTM (how does one format it, and how does one train it)

Thank you