Custom Quantization export

[MohamedA95]

Hi everyone,
I have a model that was quantized using a custom library. I would like to export it to Onnx using the Operator-oriented method. I could not find in the docs how to do it, could you provide any leads?
Thanks in advance.

[skottmckay]

I don't know what the 'operator-oriented' method is.

https://github.com/onnx/tutorials#converting-to-onnx-format lists the options for converting to onnx from various platforms.

[MohamedA95]

Hi Skottmckay, Thanks for your respond
As per the documentation, there are two ways to represent a quantized onnx model: operator- or tensor-oriented. Now I have a custom quantization flow and I want to use onnx to save the model with some extra parameters at each node.