NVIDIA’s latest “Chat With RTX” or ChatRTX, which was launched at GTC in March 2024, has received a new update. With the latest update, ChatRTX receives support for several reputed large-language AI models, which allow users to interact with their local data profoundly. ChatRTX will support Gemma, Google’s latest LLM, alongside photo and Whisper support. The ChatRTX update is part of this week’s NVIDIA AI Decoded.
NVIDIA ChatRTX uses powerful large-language AI models and harnesses the power of NVIDIA RTX-powered PCs to allow users to connect these LLMs to their data and “unleash the power of AI” on their system. The availability and interactivity of LLMs to the local data ensures the data’s security.
Also Read: Batman: Arkham Shadow Trailer Released; Coming Exclusively On Meta Quest 3
NVIDIA Adds Support For Google’s Gemma LLM And The Bilingual ChatGLM3 To ChatRTX
With the help of the retrieval-augmented generation, NVIDIA TensorRT-LLM software and NVIDIA RTX acceleration, NVIDIA brings ChatRTX's chatbot capabilities to RTX-powered Windows PCs and workstations. Thanks to already supported LLMs and the addition of Gemma, the latest open, local LLM trained by Google and ChatGLM3, an open, bilingual (English and Chinese) LLM, users can generate quick responses on their systems. With the amalgamation of powerful LLMs and NVIDIA RTX acceleration, ChatRTX can generate relevant responses to all the prompts locally.
NVIDIA has also brought in OpenAI’s Contrastive Language-Image Pre-training (CLIP), which will allow users to interact with the images present on their system without the need for complex metadata labelling. Alongside CLIP, ChatRTX will also come with an automatic speech recognition system, Whisper, which can take voice queries.
CLIP is a neural network that, through training and refinement, learns visual concepts from natural language supervision — that is, the model recognizes what it’s “seeing” in image collections.
Another update during the AI Decoded was the addition of DLSS 3.5 with Ray Reconstruction in the RTX Remix beta. RTX Remix uses AI-accelerated tools on the NVIDIA Omniverse platform, which allows modders to transform PC games into RTX remasters. Thanks to DLSS 3.5 with Ray Reconstruction, modders can improve the fidelity, responsiveness and quality of ray-traced effects of their remastered games.
AI powers other elements of the Remix workflow, too. Modders can use generative AI texture tools to analyze low-resolution textures from classic games, generate physically accurate materials — including normal and roughness maps — and upscale the resolution by up to 4x. Tools like this also save modders time, quickly handling a task that could otherwise become tedious.