Level of support for Transformer ML Models in torch MLIR

Hello all,

I wanted to find out the level of support already built for supporting Transformer ML models in Torch MLIR. Is there a web site that lists which all of Transformer models (listed on ๐Ÿค— Transformers) have been tested to be handled correctly for PyTorch Torchscript to Torch MLIR (or TOSA MLIR) conversion and subsequent lowering to Reference MLIR CPU runner ?

Thanks for any guidance on this.


Here is a list of models that Nod has been able to run using Torch-MLIR: SHARK/README.md at main ยท nod-ai/SHARK ยท GitHub

1 Like