Nvidia Unveils Latest Tech Update for ChatRTX AI Chat System

A new update for Nvidia ChatRTX, the visual computing platform designed for real-time, cinematic-quality rendering for game developers, is now available for download. This update enhances the tech demo's capabilities and expands the range of supported large language models (LLMs) for RTX-enabled AI applications.

First introduced at GTC 2024 earlier this March, the update brings several new features. ChatRTX now supports a broader spectrum of LLMs, including Gemma, Google's latest model, and ChatGLM3, an open bilingual LLM that supports both English and Chinese, offering users greater flexibility.

New photo support allows ChatRTX users to effortlessly search and interact with their photo data locally, eliminating the need for complicated metadata labeling. This is made possible through OpenAI’s Contrastive Language-Image Pre-training (CLIP).

Furthermore, users can now engage with their data verbally, thanks to the integration of Whisper, an automated speech recognition system that enhances ChatRTX's ability to understand spoken language.

ChatRTX empowers users to harness the full capabilities of AI on their RTX-enabled PCs, utilizing the accelerated performance of TensorRT-LLM software and NVIDIA RTX—all while ensuring data security through local processing. Additionally, it is available as a free reference project on GitHub, allowing developers to explore and expand AI applications using RAG technology for varied use cases.

For further details, visit the embargoed AI Decoded blog, which features a new update for the RTX Remix beta, now incorporating DLSS 3.5 with Ray Reconstruction.