Image source: NVIDIA.
NVIDIA announced the launch of its new NVIDIA AI Foundry service to help enterprises bring their custom gen AI dreams closer to fruition. NVIDIA’s new software stack is meant to ease the development and deployment of customised AI without enterprises needing to break the bank to do so.
Using semiconductor company TSMC as an example, Kari Briski, vice president of AI Software at NVIDIA, said that like TSMC, NVIDIA’s AI Foundry provides the infrastructure, libraries, tooling, compute (using DGX Cloud), and methodology for other companies to develop and customise AI models.
Customers can use AI Foundry to customise NVIDIA and open community models, including the new Llama 3.1 collection, as well as NVIDIA Nemotron, CodeGemma by Google DeepMind, CodeLlama, Gemma by Google DeepMind, Mistral, Mixtral, Phi-3, StarCoder2, amongst others.
NVIDIA DGX Cloud is a network of accelerated compute resources co-engineered with Amazon Web Services, Google Cloud, and Oracle Cloud Infrastructure. Using DGX Cloud, AI Foundry customers can develop and fine-tune custom generative AI applications easily and efficiently, and scale them out as needed without needing significant upfront investments in hardware.
Partners have already been rolling out solutions based on NVIDIA AI Foundry.
For example, Accenture unveiled its NVIDIA AI Foundry built, Accenture AI Refinery framework. This enables Accenture customers to build custom LLM models based on the Llama 3.1 collection of openly available models. Accenture customers can develop and deploy gen AI solutions that reflect their own unique culture, language, and industry.
Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.