How to export a custom fine-tuned Sortformer model for Streaming inference in Riva?

AI & Data Science Deep Learning (Training & Inference) Riva
1

Hardware - GPU A30
Hardware - CPU intel
Operating System - Ubuntu 22.04
Riva - nemo2riva: 2.22.0
I have fine-tuned a Streaming Sortformer model using NVIDIA NeMo for a speech diarization task. I now want to deploy this model using NVIDIA Riva via nemo2riva tool.
I used the command: nemo2riva --out my_model.riva my_model.nemo
But i had an error: “TypeError: SortformerEncLabelModel.forward_for_export() missing 4 required positional arguments: ‘spkache’, ‘spkache_lengths’, ‘fifo’, and ‘fìfo_lengths’
Does the current Riva version support the Sortformer architecture for streaming inference?