Siavash Nazari

I think we can close this issue. To get the custom Ops picked up by Torch-MLIR, we needed to modify QPyTorch (like [this](https://github.com/Tiiiger/QPyTorch/commit/2b5ba238734867bc5b89e838cce4ab3502a5647c)). @silvasean - I wonder how does the...

The idea of a robust conversions between ONNX and Torch front-ends seems interesting. Excited to see how this RFC evolves. It seems reasonable to avoid duplicating the effort to support...