model2.initialize_from_another_model() does not honor device='cuda'

[mw66]

    # model is trained with device='cuda', but when we do:
    model2.initialize_from_another_model(model, dataset['train_input']);

it error out:

RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!

[EladWarshawsky]

same issue for me, have to train on cpu, i'd use efficientkan for this. main thing is that the library doesn't have symbolic regression implemented yet.