Onnx support for the ICNNs

[malteal]

Hi,

I have a question about onnx support for the jax-version of ICNNs

Does anyone know of support or has tried to export the jax-version of the ICNN to onnx with the gradient of the network being the output of the onnx model?

I have a Pytorch-version of ICNNs that can be exported to onnx using onnxruntime-training, however, I have not been able to do it using the ott framework.

Best regrads,
Malte