RuntimeError: svd_cuda

[seeyourmind]

Hi, when I ran the Pytorch example, I get the error as follows:

File "/LIDVAE/sequential_pytorch/src/modules/decoders/dec_lstm.py", line 74, in forward
    U, S, V = torch.svd(self.weight+1e-5*torch.eye(n=self.weight.shape[0], m=self.weight.shape[1]).cuda())
RuntimeError: svd_cuda: (Batch element 0): The algorithm failed to converge because the input matrix is ill-conditioned or has too many repeated singular values (error code: 1057).

Is there any solution for this error? Thanks a lot!