Skip to content

Conversation

@houseroad
Copy link
Member

Add the symbolic for UpsamplingNearest2d.

Locally tested on test_operators.py and test_caffe2.py.

Will update ONNX and ONNX-Caffe2 later.

@houseroad houseroad force-pushed the upsamplingnearest2d branch from c874a9a to 2a517c0 Compare November 3, 2017 04:19

This comment was marked as off-topic.

This comment was marked as off-topic.

ezyang
ezyang previously requested changes Nov 4, 2017
Copy link
Contributor

@ezyang ezyang left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Use the helper function here. Also, we should export this as a Caffe2 vendor extension op while it is not supported in ONNX (talk to me about how to do this.)

@ezyang ezyang added the module: onnx Related to torch.onnx label Nov 4, 2017
@houseroad
Copy link
Member Author

Just realize that this one is a little bit tricky, since ATen has no corresponding implementation.

@houseroad
Copy link
Member Author

onnx/onnx#180, now we have it as an experimental op. Later we may polish it. I think upsample op is useful/necessary.

@pytorchbot
Copy link
Collaborator

Build finished.

1 similar comment
@pytorchbot
Copy link
Collaborator

Build finished.

@houseroad
Copy link
Member Author

I think Dima already updated the symbolic. But it got changed in ONNX, so I will update this PR to fix the problem.

@ezyang ezyang merged commit 60d86ba into pytorch:master Nov 29, 2017
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

module: onnx Related to torch.onnx

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants