Dilshod Tadjibaev

Yes, I think it'd be useful. So basically it would mean to implement a new backend that can use wasi-nn apis and have the code compiled to wasm32? Linking wasi-nn...

It would worth coming back to this when `wasi-nn` matures and becomes generally available on major cloud services. I do see benefits deploying models to the edge compute easily. However,...

Thanks for filing. Definitely useful to support Gemm ONNX operator. However, you mentioned it's useful for optimization, what do you mean? So you know, Burn does not have native Gemm...

Closing this since GEMM is supported OP via #2841

We will be happy to accept your contribution. You can launch contributor book to read up more on burn onnx. You should try the main branch. We have over 20...

> @doronnac Have you made any progress on this, or at least come up with a temporary fix/workaround? I am also running into this issue. Please try main branch if...

> @antimora Is there a particular reason for not supporting rank > 1 for the indices in `dim_inference` for `Gather` nodes? It seems there is already code in place to...

Looks like the upstream issue is fixed. Do we need to do anything @LaurentMazare , @syl20bnr ?

A new issue should be filed with Candle and come back to update Candle backend.

We don't need to track the output shapes statically because the Burn API tracks and calculates dynamically. Burn needs only rank output. However, onnx-ir might have a feature that tracks...