Hi,
Briefly, I’ve been able to Install CUDA Toolkit on Windows and WSL2, and managed to update the CUDA Toolkit v. 12.8. So far, so good. Also have Nvidia GeForce RTX, so compatible hardware.
Now looking for a little guidance regarding how to use the NIMS, NEMO, (2-D & 3-D avatar chat assistants, multi-modal RAG, etcetera) in a hybrid or full cloud deployment. I understand that I will have to link or access my server provider account (AWS) though I am unsure the process with the Docker containers, do I require the API Key as well, is this an internal API Key for Nvidia? And is there one more step I need to do to have access or install a requirement of Nvidia images/ containers?
I am familiar with Docker, and have a build that abstracts much of the CLI for running containers, using four main inputs, Container Name, Image Name, * Container Port: “8888/tcp” → Host Port: “8888” ; Environment Variables (Optional), which then allows me to login to this above example at " login at [http://localhost:8888]. (Screenshot attached of the Universal form, under "Integrations; LHS/LHS)
Only pointing this out as I am somewhat familiar with Docker and containers, though as I’m used to using my build to pull images and run containers,
do I also need the API Key an an internal key to pull the Nvidia images and run the containers, and do I need to do something prior to this. A 24 hour crash course in the Nvidia suite has only got me so far.
I look forward to any and all assistance, end goal is how to make use of the NIMS, NEMO, and deploy them for clients, do I setup local and cloud (hybrid) them deploy, do I “build” in the cloud. Too many questions, I will await some sound advice.
Many thanks.
Kind regards,
Andrew