Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Fix torch.tensor(...) device-type calculation when used with numpy an…#6995

Merged
gchanan merged 5 commits into
pytorch:masterfrom
gchanan:nparray_cuda
Apr 27, 2018
Merged

Fix torch.tensor(...) device-type calculation when used with numpy an…#6995
gchanan merged 5 commits into
pytorch:masterfrom
gchanan:nparray_cuda

Conversation

@gchanan
Copy link
Copy Markdown
Contributor

@gchanan gchanan commented Apr 26, 2018

…d type inference.

@gchanan gchanan requested a review from colesbury April 26, 2018 15:14
@gchanan
Copy link
Copy Markdown
Contributor Author

gchanan commented Apr 26, 2018

Should fix #6985.

@gchanan gchanan mentioned this pull request Apr 26, 2018
Closed
@zou3519
Copy link
Copy Markdown
Contributor

zou3519 commented Apr 26, 2018 •
edited
Loading

@gchanan The same problem exists for torch.tensor with a CPU tensor:


In [6]: import torch

In [7]: x = torch.randn(3, 3, requires_grad=True)

In [8]: torch.tensor(x, device='cuda')
Out[8]:
tensor([[-0.3721,  0.5129,  0.6696],
        [-0.8509,  0.4349,  1.1458],
        [ 0.7075, -3.0103, -0.6825]])

In [9]: torch.tensor(x, device='cuda').device
Out[9]: device(type='cpu')

I think this line

pytorch/torch/csrc/utils/tensor_new.cpp

Line 183 in 2b44c42

const auto& type_to_use = type_inference ? var.type() : type;
might need to be changed too.

@gchanan
Copy link
Copy Markdown
Contributor Author

gchanan commented Apr 26, 2018

@zou3519 is correct; I'll fix that before merging.

@onnxbot-worker-3 onnxbot-worker-3 mentioned this pull request Apr 26, 2018
Closed
@karandwivedi42
Copy link
Copy Markdown
Contributor

karandwivedi42 commented Apr 26, 2018

@gchanan Can you please briefly explain how this fixed the problem?

type_inference ? type.toScalarType(var.type().scalarType()) : type;

@gchanan
Copy link
Copy Markdown
Contributor Author

gchanan commented Apr 26, 2018

@karandwivedi42 type inference means in this case means dtype (scalar type) inference, while the ATen type holds (scalar type, layout, device type) e.g. CUDASparseFloatType. So, if we do type inference, we only want to use the scalar type of the incoming variable, not the device-type / layout as well.

@gchanan
Copy link
Copy Markdown
Contributor Author

gchanan commented Apr 26, 2018

Actually, the issue with Variable is more complicated; you want to do type-inference on cuda-ness if device isn't specified, otherwise you don't.

@gchanan gchanan merged commit 361648a into pytorch:master Apr 27, 2018
Jorghi12 pushed a commit to wsttiger/pytorch that referenced this pull request May 10, 2018
@gchanan @Jorghi12
Fix torch.tensor(...) device-type calculation when used with numpy an… (
4917a27 
pytorch#6995)

* Fix torch.tensor(...) device-type calculation when used with numpy and type inference.

* Fix tensor device type inference as well.

* Better variable type inference: infer cuda-ness only if device is not specified.
weiyangfb pushed a commit to weiyangfb/pytorch that referenced this pull request Jun 11, 2018
@gchanan @weiyangfb
Fix torch.tensor(...) device-type calculation when used with numpy an… (
15da4e1 
pytorch#6995)

* Fix torch.tensor(...) device-type calculation when used with numpy and type inference.

* Fix tensor device type inference as well.

* Better variable type inference: infer cuda-ness only if device is not specified.
@zou3519 zou3519 mentioned this pull request Jul 17, 2018
Closed
laurentdupin pushed a commit to laurentdupin/pytorch that referenced this pull request Apr 24, 2026
@gchanan
Fix torch.tensor(...) device-type calculation when used with numpy an… (
2d9ab2c 
pytorch#6995)

* Fix torch.tensor(...) device-type calculation when used with numpy and type inference.

* Fix tensor device type inference as well.

* Better variable type inference: infer cuda-ness only if device is not specified.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@ezyang ezyang ezyang approved these changes

@colesbury colesbury Awaiting requested review from colesbury

@apaszke apaszke Awaiting requested review from apaszke

@soumith soumith Awaiting requested review from soumith

@zdevito zdevito Awaiting requested review from zdevito

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

4 participants

@gchanan @zou3519 @karandwivedi42 @ezyang