Llama-3.1-Nemotron-Ultra-253B-v1 API not working

AI & Data Science NVIDIA NIM Models
1

It’s been saying ‘Internal server error’ since Saturday. Can we look into this? Thanks.

2

Thanks for bringing this to our attention - I’ve asked the team to redeploy the endpoint.

Sophie

1 Like
3

It’s now back up and running - thanks again for letting us know it wasn’t working @vutran_kpt

Sophie

1 Like
4

Thank you for fixing it. Can we also look into Llama-3.3-Nemotron-Super-49B-v1.5? If we disable reasoning, the model shows ‘ErrorEngineCore encountered an issue. See stack trace (above) for the root cause.’ Thanks.

5

Hi @vutran_kpt,

I’m struggling to recreate the error you’re seeing with the llama-3_3-nemotron-super-49b-v1_5 model - are you using it through the web UI or deploying elsewhere?

Thanks,

Sophie

6

image
image1800×882 58 KB

The Web UI shows this error when using the model with reasoning off. Same with API.

7

Interesting! I was asking the model the suggested math question, rather than ‘counting the r’s in strawberry’ question - it works fine on the maths!

I’ll let the team know and get it sorted!

Sophie