I am trying to get familar with NIM Pod deployment process and test the embedding & chat model. So in a Ec2 Box (with ubuntu 22) I installed the ngc.
I have an nvidia develoepr account. There I created a personal key (Url is : Dashboard | NVIDIA NGC) with permissions nv-cloud-functions, artifact-catalog & secrets-manager (Only these three options are available there.
After this I tried to execute ngc config and added the key there. Then i tried to download the nim-cli using following command
ngc registry resource download-version nvidian/nim-llm-dev/nim-cli
But I am always getting authentication error. “403 Response: Access Denied” Please look into this. The Key ID is c0…176 and key-name is ngc-access-1".
What am I missing here? What should I do? I want to run nvidia/nim/nim-embedding-e5-small and nvidia/nim/nim-gpt-2
Hi @dwaipayan.basu we are looking into this. will get back to you asap. thanks for your patience.
Hi @dwaipayan.basu a quick update - the registrynvidian/nim-llm-dev/nim-cli is only accessible to those who work at NVIDIA. Thanks to your post we realised that this registry made its way into some documentation - we will update that as soon as possible.
With regards to finding models you want to use please follow the instructions here Getting Started — NVIDIA NIM for Large Language Models (LLMs) to launch with Docker.
You mention two models - nvidia/nim/nim-embedding-e5-small and nvidia/nim/nim-gpt-2. Neither of these are the addresses of models in our registry. Perhaps you mean nv-embedqa-e5-v5 Model by NVIDIA | NVIDIA NIM for the embedding model? Not sure about the gpt-2 model… To explore the full model catalog, visit build.nvidia.com.
Thanks again for bringing the incorrect documentation regarding nvidian/nim-llm-dev/nim-cli to our attention. And do reach out if you have follow up questions.
Best,
Sophie