ONNX export to triton kernel

[edgarriba]

Hi, I was wondering if any one tried to export/generate an entire ONNX model as a single triton kernel ? I would like to be able to run entire models using triton kernels in rust as in this example here https://github.com/jeromeku/triton-rs/tree/main

Osidge question, what’s the progress on the cpu backend — is there any timeline to be merged into main ? Are you planning to support other backends too ?