How to deploy custom models to AWS EKS

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

Please provide the following information when requesting support.

Hardware - GPU g5 instance
Hardware - CPU g5 instance
Operating System
Riva Version 2.10
TLT Version
How to reproduce the issue ?

I would like to understand more the deploy process related to custom rmir model file.

I’m trying to deploy AWS EKS with my custom rmir files but I’m a bit confused with this:

#for non ngc based images or custom images.
raw_server_image: “”
raw_servicemaker_image: “”

¿What to put in raw vars and where I change to my custom rmir files?

I got it working locally, I’m just a bit lost deploying to EKS. ¿Any help? Thanks.

2

Hi @carlfm01

Thanks for your interest in Riva,

I will check with the team and provide updates

Thanks

3

Hello ¿Any updates?

4

Hi @carlfm01

My Sincere Apologies for the delay

Kindly use the existing AWS EKS tutorial - tutorials/deploy-eks.md at main · nvidia-riva/tutorials · GitHub

  1. What to put in raw vars and where I change to my custom rmir files ?
    ANS - Use field: ngcModelConfigs in Helmchart → Values.yaml

Thanks