Nvidia unveiled a collection of cloud services at its annual GPU Technologies Conference to assist companies in creating and running generative AI models trained on unique data and tailored for “domain-specific tasks,” such as producing advertising text.
Each particular cloud service offered by Nvidia AI Foundations, such as Nvidia NeMo for language models and Nvidia Picasso for the creation of images, videos, and 3D material, contains pre-trained models, data processing frameworks, APIs, and support from Nvidia engineering staff. Enterprises can use Nvidia’s cloud or another service to execute models once they are prepared for deployment.
The NeMo generative AI cloud service is currently available in early access. There is a private preview of the Picasso service. Applications must be submitted in order to participate in either.