Loading custom n-dimensional data from memory

I am building a CNN classifier and struggling with preparing my data for the model.

My training data is inside a list train_ds of length N. Each item in this list is of type numpy.ndarray with a shape of (7, 200, 200). The labels for my data are in a separate list of the same length. Another separate list for validation val_ds is also available.

As you can see, my data is already loaded into memory and I don’t want to do any data transformation or augmentation. How should I go about loading the data into a learner? is there any simpler way other than creating my own ItemList class ?