Please clarify how to use recognition API on customized RMIR

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

Please provide the following information when requesting support.

Hardware - GPU T4
Hardware - CPU
Operating System Ubuntu 20.04
Riva Version 2.5.0
TLT Version (if relevant)

I successfully ran quickstart riva_install for the conformer model (also modified the config file for different models and they all worked as expected).
I used the basic python script that calls riva.client package.

I’m trying to modify the language to use en-GB (hopefully in addition to and not instead of) the en-US - but I’m not understanding all the steps needed to do this. I got some help from an Nvidia alliance manager who led me to successfully build an rmir using riva-build as well as create the model using riva-deploy (inside the servicemaker 2.5.0 docker)

I don’t know how to use the models/rmir I created in place of (or ideally in addtion to) the default rmirs/models of en-US conformer so when I run riva.client in my python script it knows to use the new model.

Can anyone help?
Thanks
edmond@magnifi.ai

I want to know if/how to use th

2

Hi @edmond4

Thanks for your interest in Riva,

I will check with the team and get back

Thanks

3
1 Like
4

Hi @edmond4

Please find the documentation in below link for customized rmir usage
https://docs.nvidia.com/deeplearning/riva/user-guide/docs/model-overview.html#deploying-your-custom-model-into-riva

Thanks