Add support for named outputs in MLF archive#13704
Conversation
|
Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from Reviewers by @-ing them in a comment.
Generated by tvm-bot |
Following from apache#12789, this adds support for determining the output tensor name from the input model within the MLF metadata json. Co-authored-by: Ashutosh Parkhi <ashutosh.parkhi@arm.com>
Mousius
force-pushed
the
mlf-named-metadata
branch
from
fc5aec8 to
b32beba
Compare
mehrdadh
left a comment
mehrdadh
left a comment
There was a problem hiding this comment.
I don't think this requires an RFC, but could you please add some details of the metadata.json new structure to the PR description. Mainly explain how would someone access the output/input names for better visibility.
|
|
||
|
|
||
| if __name__ == "__main__": | ||
| sys.exit(pytest.main([__file__] + sys.argv[1:])) |
There was a problem hiding this comment.
Please update this to tvm.testing.main()
There was a problem hiding this comment.
This feels like it'd be better done as a blanket change rather than polluting many individual patches?
There was a problem hiding this comment.
either way is fine. We don't have a lint step to catch these, that's why I've trying to notify people to fix it as I see them in the PR.
There was a problem hiding this comment.
I'd rather not introduce unrelated changes into my patch, I can try a mass find and replace in a future patch? 😸
Hi @mehrdadh, #12789 adds the functionality in the metadata, this adds support for output names - I haven't changed the structure. |
|
@tvm-bot rerun |
|
@ashutosh-arm I think we still need #13655 for the checking of the actual constant which was broken before? this just checks the JSON |
|
fyi @Mousius @mehrdadh this seems to have merge conflicted with something once it was merged in 329584b so CI is now failing, I have #13739 up to revert it, we can wait and see if that fixes the failures we're seeing on example failure: https://ci.tlcpack.ai/blue/organizations/jenkins/tvm-cortexm/detail/main/115/pipeline |
This PR fixes error introduced by #13704
Following from apache#12789, this adds support for determining the output tensor name from the input model within the MLF metadata json. Co-authored-by: Ashutosh Parkhi <ashutosh.parkhi@arm.com>
This PR fixes error introduced by apache#13704
5 participants
Following from #12789, this adds support for determining the output tensor name from the input model within the MLF metadata json.
Co-authored-by: Ashutosh Parkhi ashutosh.parkhi@arm.com