The name of ONNF is misleading about onnx-mlir

[doru1004]

Fixed by #32

[tjingrant]

If we go with the onnx-mlir branding, will it resolve your concern? It'd be a bit simpler.

[caoimhinuibrian]

I agree. We should just drop the ONNF name, it’s not actually adding anything. Kevin

…

On Mar 1, 2020, at 5:27 PM, Tian Jin ***@***.***> wrote:  If we go with the onnx-mlir branding, will it resolve your concern? It'd be a bit simpler. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[tjingrant]

Sounds good, will switch to onnx-mlir throughout our codebase.
@linkerzhang please let me know your thought!

[linkerzhang]

Thanks! onnx-mlir is better than onnf :), though I'll be open to other even more descriptive names if any (I'm not good at naming at all, btw.).

[bhack]

As the ONNX dialect is a quite different third party dialect cause It is related to a standard do you plan to contribute It in https://github.com/llvm/llvm-project/tree/master/mlir/lib/Dialect?

[bhack]

There is just a new doc if you are interested https://github.com/llvm/llvm-project/blob/master/mlir/docs/CreatingADialect.md

[tjingrant]

It is one of our goals to try and see if there is a place for ONNX to exist within MLIR core. But it seems like we have some distance to cover and more work/discussions are needed before we can tell whether MLIR core is the right place for ONNX dialect and whether the MLIR community is willing to accept the ONNX dialect - c.f. https://groups.google.com/a/tensorflow.org/forum/#!msg/mlir/CUL3_8fnJb4/0Uvt3M-FAQAJ.

[bhack]

Yes but the MLIR community position was a little bit evolved from the google group post.
I.e. See https://llvm.discourse.group/t/development-of-high-level-tensor-compute-primitives-dialect-s-and-transformations/388/57

[AlexandreEichenberger]

Thanks, nice to hear that they are softening on the usefulness of the ONNX dialect. I concur with Tian that it would be best for us to continue developing the dialect separately for a little while so that we may come with a comprehensive approach. Alexandre-----------------------------------------------------------------------------------------------------Alexandre Eichenberger, Principal RSM, Advanced Compiler Technologies- research: compiler optimization (OpenMP, GPU, SIMD)- info: [email protected] http://www.research.ibm.com/people/a/alexe- phone: [phone] (work), 914-312-3618 (cell)     ----- Original message -----From: bhack <[email protected]>To: onnx/onnx-mlir <[email protected]>Cc: Subscribed <[email protected]>Subject: [EXTERNAL] Re: [onnx/onnx-mlir] The name of ONNF is misleading (#4)Date: Sat, Mar 7, 2020 10:50 AM  Yes but the MLIR community position was a little bit evolved from the google group post.I.e. See https://llvm.discourse.group/t/development-of-high-level-tensor-compute-primitives-dialect-s-and-transformations/388/57 —You are receiving this because you are subscribed to this thread.Reply to this email directly, view it on GitHub, or unsubscribe.  

[bhack]

Ok but try to early interact with them on discuss/Discord.