Can Torchdiffeq add support for sparse tensors for y0

[032004129xuzhiyong]

I want a very large sparse tensor as an extra parameter. As suggested by the previous question, I put this as part of the status, but it reported an error. Can Torchdiffeq add support for sparse tensors for y0?

(func, y0, t, rtol, atol, method, options, event_fn, SOLVERS)
    218 rtol = _tuple_tol('rtol', rtol, shapes)
    219 atol = _tuple_tol('atol', atol, shapes)
--> 220 y0 = torch.cat([y0_.reshape(-1) for y0_ in y0])
    221 func = _TupleFunc(func, shapes)
    222 if event_fn is not None:

RuntimeError: reshape is not implemented for sparse tensors