From NVIDIA: 2024-10-02 09:00:02

AI technology is being integrated into various applications to enhance user experiences and increase efficiency, including web browsing. Brave’s AI assistant, Leo AI, can summarize articles, answer questions, and more. The technology behind Brave and other AI tools combines hardware, libraries, and ecosystem software optimized for AI needs, powered by NVIDIA GPUs.

NVIDIA GPUs, with Tensor Cores designed for AI acceleration, are crucial for running AI applications efficiently. Software like inference libraries, such as llama.cpp, plays a critical role in optimizing AI tasks for hardware to run. Ollama, an open-source project, provides access to llama.cpp features, supporting various AI models and delivering local AI capabilities.

Brave’s Leo AI can run in the cloud or locally through Ollama, offering privacy and constant availability. Running AI locally eliminates the need for unrestricted cloud access and allows for a wider variety of open-source models. With RTX, users can expect faster responses to questions and content requests when using Brave with Leo AI and Ollama.

Installing Ollama is simple, allowing users to download supported models and interact with the local model from the command line. By configuring Leo AI to use the locally hosted model through Ollama, users can seamlessly switch between cloud and local models. Developers can explore Ollama and llama.cpp in the NVIDIA Technical Blog for more information.



Read more at NVIDIA: Leo AI and Ollama Bring RTX Local LLMs to Brave Browser