Skip to content

Use the Java Op generator #287

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

Merged
merged 2 commits into from
Apr 20, 2021
Merged

Conversation

rnett
Copy link
Contributor

@rnett rnett commented Apr 11, 2021

Enables the Java op generator during install and removes the C++ generator. Builds fine locally, although there are still Javadoc errors in core with openjdk14 for non-opgen classes (i.e. scope param tags are used in the Ops class). I'll leave them to another PR.

@karllessard can you flag this for CI build?

It also fixes #286.

@karllessard
Copy link
Collaborator

Ok so it passes a few platforms in the full build but timed out eventually and Bazel-caching is not enabled for PRs so we will now give it a shot in master. I did not went through all the 1000+ regenerated ops to find issues but we'll fix them as we find them, if any.

Great work @rnett ! I might miss that old C++ code of mine but that is just me being sentimental :)

@karllessard karllessard merged commit 7d167ad into tensorflow:master Apr 20, 2021
@karllessard
Copy link
Collaborator

karllessard commented Apr 21, 2021

@rnett can you please check why the Windows build is failing after merging this PR? It seems to be related.

@karllessard
Copy link
Collaborator

@rnett , we've noticed that on Windows, looks like there are only 603 files being compiled while on Linux, there are 1807. If that can give some cue...

@rnett
Copy link
Contributor Author

rnett commented Apr 21, 2021

Odd, that seems like it's probably a problem with the op export code rather than the op generation. Will look.

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

Successfully merging this pull request may close these issues.

The protobuf version doesn't match the underlying tensorflow version
2 participants