Wenbing Li

> For my own case, I would love to be able to have my data preprocessing directly available in Triton, such as tokenization. I can do it locally if I...

check there for for a complete GPT-2 model inference example https://github.com/microsoft/onnxruntime-extensions/blob/main/tutorials/gpt2bs.py

this is too general, and looks it is not relevant to this repo.

Thanks for the reporting, @turneram The model earlier than opset 7 were lack of maintenance, and I suppose these would be removed in the future.

@YuhengHuang42 , Thanks for reporting the issue. These model most are legal while ONNXRuntime doesn't support some old opset for sake of its binary size. So actually it's ONNXRuntime issue....

checked the PR for a end-to-end GTP-2 model: https://github.com/onnx/models/pull/445

what's that? an onnx to tensorrt conversion tool?

this looks tensorflow model, can you try this tool https://github.com/onnx/tensorflow-onnx to convert it to ONNX model?

I didn't see any pre/post process steps for Alexnet model, https://github.com/onnx/models/tree/master/vision/classification/alexnet . Can you share more information about the pre/post process?

check this model for an example to see how you can decode the output. https://github.com/onnx/models/tree/master/text/machine_comprehension/gpt2-bs