Vmap: inplace arithmetic error

1

Hello everyone, I’m struggling to understand why vmap is not working, attached you can find a minimal code that reproduces the error:

import torch

def dummy_func(amount):
    test = torch.eye(4,4)
    test[2,3] = amount
    return test

if __name__ == "__main__":
    test_input = torch.tensor([1., 2., 3., 4., 5.])

    test_output = torch.vmap(dummy_func)(test_input)

The error you should get is the following:

RuntimeError: vmap: inplace arithmetic(self, *extra_args) is not possible because there exists a Tensor `other` in extra_args that has more elements than `self`. This happened due to `other` being vmapped over but `self` not being vmapped over in a vmap. Please try to use out-of-place operators instead of inplace arithmetic. If said operator is being called inside the PyTorch framework, please file a bug report instead.

Any help would be very appreciated! :slightly_smiling_face:

2

The error you’re getting is complaining about please try to use out-of-place operators

So you can’t have a write to an existing tensor

3

Thanks for your reply! Do you have a fix to suggest?