How do you train models when your training data is a 3D tensor?

I'm working with data collected from implanted electrode arrays collected from many patients, and have successfully trained models for individual patients, but I would like to train a model using all the patients' data. I've gotten stuck on a task since my data is size num\_users x num\_features x num\_datapoints, and I'm not sure how to train a machine learning model on this data. I can think of a few ways to do so but all appear to introduce undesirable …

arrays data deeplearning patients tensor train training training data