RuntimeError: Expected object of scalar type Double but got scalar type Float for argument #2

Question

I have a PyTorch LSTM model and my forward function looks like:

    def forward(self, x, hidden):
        print('in forward', x.dtype, hidden[0].dtype, hidden[1].dtype)
        lstm_out, hidden = self.lstm(x, hidden)
        return lstm_out, hidden

All of the print statements show torch.float64, which I believe is a double. So then why am I getting this issue?

I've cast to double in all of the relevant places already.

MogaGennis · Accepted Answer · 2021-05-03 13:19:02Z

6

Make sure both your data and model are in dtype double.

For the model:

net = net.double()

For the data:

net(x.double())

It has been discussed on PyTorch forum.

edited May 3, 2021 at 13:19

MogaGennis

1172 silver badges8 bronze badges

answered Mar 2, 2020 at 19:14

David Ferenczy Rogožan

25.7k12 gold badges87 silver badges73 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

RuntimeError: Expected object of scalar type Double but got scalar type Float for argument #2

1 Answer 1

Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Related