A question about new feature SparseGaussianAdam

[rylynchen]

Why SparseGaussianAdam do not support group["params"] multi elements, such us torch.optim.Adam ?

class SparseGaussianAdam(torch.optim.Adam):
    def __init__(self, params, lr, eps):
        super().__init__(params=params, lr=lr, eps=eps)
    
    @torch.no_grad()
    def step(self, visibility, N):
        for group in self.param_groups:
            lr = group["lr"]
            eps = group["eps"]

            assert len(group["params"]) == 1, "more than one tensor in group"
            param = group["params"][0]
            if param.grad is None:
                continue
        ....

class Adam(Optimizer):
    @_use_grad_for_differentiable
    def step(self, closure=None, *, grad_scaler=None):
        ...
            for p in group['params']:
                if p.grad is not None:
                    params_with_grad.append(p)
                    if p.grad.is_sparse:
                        raise RuntimeError('Adam does not support sparse gradients, please consider SparseAdam instead')
                    grads.append(p.grad)
                    ...
                    state_steps.append(state['step'])