NVIDIA accelerates development of custom generative AI models for enterprises

Developers can experiment with the new NVIDIA AI Foundation models directly from the browser, test in their applications using NVIDIA AI Foundation Endpoints, and then customize them using their unique business data.

Today's landscape of free, open source large language models (LLMs) is like a smorgasbord for enterprises. This abundance can be overwhelming for developers building their own generative AI applications, as they have to navigate unique project and business requirements, including compatibility, security, and the data used to train the models.

NVIDIA AI Foundation Models, a collection of enterprise-grade pre-trained models, gives developers a head start on bringing generative AI to enterprise applications.

NVIDIA-optimized Foundation models accelerate innovation

NVIDIA AI Foundation models can be used through a simple user interface or API, directly from a browser. In addition, these models can be accessed through NVIDIA AI Foundation Endpoints to test model performance in enterprise applications.

Available models include leading community models such as Llama 2, Stable Diffusion XL, and Mistral, which are formatted to help developers simplify customization with their own data. In addition, the models have been optimized with NVIDIA TensorRT-LLM to provide the highest throughput and lowest latency, and to run at scale on any NVIDIA GPU-accelerated stack. For example, the Llama 2 model optimized with TensorRT-LLM runs nearly 2x faster on NVIDIA H100.

The new NVIDIA Nemotron-3 8B base model family supports the creation of the most advanced enterprise chat and Q&A applications for a wide range of industries, including healthcare, telecommunications, and financial services.

The models are the starting point for building secure, production-ready generative AI applications, are trained on responsive datasets, and operate with performance comparable to much larger models. This makes them ideal for enterprise deployments.

Multilingual capabilities are a key differentiator of the Nemotron-3 8B models. Out of the box, these models support over 50 languages, including English, German, Russian, Spanish, French, Japanese, Chinese, Korean, Italian and Dutch.

Accelerated setup and deployment

Enterprises using generative AI in business functions need AI foundry to customize models for their unique applications. NVIDIA's AI foundry comprises three elements - NVIDIA AI Foundation Models, the NVIDIA NeMo framework and tools, and NVIDIA DGX Cloud AI supercomputing services. Together, they provide a comprehensive enterprise offering for creating custom generative AI models.

Importantly, enterprises own their customized models and can deploy them virtually anywhere on accelerated computing with security, stability, and enterprise-grade support using NVIDIA AI Enterprise software.

NVIDIA AI Foundation models are freely available for experimentation in the NVIDIA NGC and Hugging Face catalog, and are also hosted in the Microsoft Azure AI model catalog.