ChatRTX, a demo app from NVIDIA that allows users to personalize a GPT large language model (LLM) connected to a variety of content, including documents, notes, images, or other data, has been improved courtesy of a new update that delivers support for a handful of AI models, including new features, NVIDIA has announced. NVIDIA ChatRTX supports RTX-powered Windows PCs and workstations, or, more specifically, those with a GeForce RTX 30/40 Series GPU or RTX Ampere or Ada GPU with at least 8 GB of VRAM.
New models and features include:
- Support for additional LLMs, including Google’s Gemma
- Support for ChatGLM3, an open, bilingual (English, Chinese) LLM
- Support for Contrastive Language-Image Pre-training (CLIP) from OpenAI, allowing users to interact with image data more easily
- Support for Whisper, allowing users to chat with their data using their voice
A look at the features in action:
NVIDIA on some of inner workings of ChatRTX:
ChatRTX uses retrieval-augmented generation, NVIDIA TensorRT-LLM software and NVIDIA RTX acceleration to bring chatbot capabilities to RTX-powered Windows PCs and workstations. Backed by its powerful large language models (LLMs), users can query their notes and documents with ChatRTX, which can quickly generate relevant responses, while running locally on the user’s device.