Nvidia AI Foundations is a family of cloud services through which organizations can build their own large language models (LLM), the technologies at the heart of generative AI systems, run them at scale, and call them from enterprise applications through Nvidia’s APIs.
There will be three Foundations services at launch, still in limited access or private preview for now: NeMo for text generation, Picasso for visuals, and BioNeMo for molecular structures. Each of the three offerings will include pre-trained models, data processing frameworks, custom databases, inference engines and APIs that organizations can access from a browser, Nvidia said.
NeMo, the text-based service, includes a variety of pre-trained AI models that users can further train on their data to match them with domain-specific knowledge.
Financial data provider Morningstar is already exploring how it can use NeMo to extract useful information about markets from raw data, drawing on the expertise of its own staff to fine-tune the models.
The Picasso service will allow organizations to train models to create custom images, videos and even 3D models in the cloud. Nvidia is partnering with Adobe to provide such creative capabilities within its tools for creative professionals such as Photoshop and After Effects.
Copyright bleach
Nvidia wants to clean up its fast-and-loose graphics AI reputation with the rights of the artists and photographers whose work the models are trained on. There are concerns that using such models to create a derivative work could expose organizations to claims for copyright infringement. Nvidia hopes to allay those concerns with a licensing deal with stock image library Getty Images, which says it will pay artists royalties on revenue generated by models trained on the works in its database.
