New Nvidia-Hugging Face Collaboration Streamlines AI Model Deployment

At the SIGGRAPH 2024 conference in Denver, Nvidia announced its latest efforts to streamline AI application development. The company unveiled a new inference-as-a-service offering in collaboration with model repository Hugging Face, alongside the introduction of new microservices tailored for industrial generative AI use cases.

Hugging Face’s new service, powered by Nvidia’s DGX Cloud and inference microservices, allows developers to swiftly deploy popular large language models such as Meta’s Llama 3 and Mistral’s AI models. This service aims to accelerate the deployment process for developers working with these advanced AI tools.

Nvidia's microservices, officially named Nvidia NIM, consist of AI models served in optimized containers, enabling seamless integration with developers' applications. Initially launched in June, NIM supports over 40 models developed by Nvidia and other contributors. These microservices are available to businesses through the Nvidia AI Enterprise software suite, priced at $4,500 per GPU annually, and are free for Nvidia Developer Program members.

The new Hugging Face inference service facilitates rapid prototyping with open-source AI models hosted on the Hugging Face Hub, making it easier for developers to move from experimentation to production. This service builds on the existing Train on DGX Cloud service introduced by Hugging Face and Nvidia at SIGGRAPH 2023.

Nvidia’s latest initiatives highlight its commitment to simplifying AI development and expanding the accessibility of powerful AI tools for industrial applications.

Source: Dylan Martin / CRN