Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Adding custom operators in onnxruntime #43

Closed
HemaSowjanyaMamidi opened this issue Jan 18, 2021 · 5 comments
Closed

Adding custom operators in onnxruntime #43

HemaSowjanyaMamidi opened this issue Jan 18, 2021 · 5 comments

Comments

@HemaSowjanyaMamidi
Copy link

How to add custom operators using ort-customops(framework independent)?
I successfully added the operator in onnx. So for the model to be compiled using the runtime it should be added in runtime as well. Need help in this area. Thanks.

@wenbingl
Copy link
Member

From the name itself, "onnxruntime customops". this library cannot be framework independent.

@xadupre
Copy link
Member

xadupre commented Jan 21, 2021

What operator do you have in mind? You can look into PR #24 as an example.

@HemaSowjanyaMamidi
Copy link
Author

@xadupre I just want to test adding a new operator to ort nothing in specific. Can add any operator which even involves computation?

@xadupre
Copy link
Member

xadupre commented Jan 21, 2021

If you just want to try, you can just create a PR. I usually create two versions of the same operator, one in python, one in C++ to compare the outputs. I recommend starting with python. The current main branch only run CI on Windows, #44 does it for Linux, MacOS, Windows and has a more consistent behaviour when it comes to strings. It should be merged in a couple of days.

@GeorgeS2019
Copy link

GeorgeS2019 commented Jun 3, 2022

The feature request here: Add an entry on Build for Custom Op development could include the discussion here.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants