You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It looks like ONNX-MLIR is now a lot more mature and supports 116 ML models out of 128 Models from the ONNX Zoo, including large models like Bert and GPT. Ref: onnx/onnx-mlir#128.
We should seriously consider upgrading the onnx-mlir version being used with LLTFI.
The text was updated successfully, but these errors were encountered:
@abrahamchan I have updated the onnx-mlir version to commit id 9c62e1a4. However, I had to make certain changes in instrumentation passes of the onnx-mlir code to get LLTFI working with it. We have to decide whether 1) I should keep my changes as a patch and apply the patch to onnx-mlir? Or 2) we fork the onnx-mlir repository to DependableSystemsLab GitHub org and I make changes to the forked version of onnx-mlir.
It looks like ONNX-MLIR is now a lot more mature and supports 116 ML models out of 128 Models from the ONNX Zoo, including large models like Bert and GPT. Ref: onnx/onnx-mlir#128.
We should seriously consider upgrading the onnx-mlir version being used with LLTFI.
The text was updated successfully, but these errors were encountered: