Nvidia’s ChatRTX app released its planned version 0.3 update today on The Nvidia website. The ChatGPT-like app update unlocks a host of new features first introduced at Nvidia. CGV Conference in March, including photo search capabilities, AI-based voice recognition, and compatibility with even more LLMs.
ChatRTX represents another forward-thinking technology demonstration from Nvidia of the power of AI when powered by consumer RTX graphics cards. The ChatRTX program provides a ChatGPT-adjacent experience installed locally on your computer, with the ability to leverage your data in AI use cases without the risk of having that data stolen for training or further uses. harmful. An example of the program in practice can be seen in this example from enthusiast putting an RTX card into a NAS to use ChatRTX’s data synthesis capabilities on their network-wide data.
We were able to download the new version before its official release and check out some of the new features. ChatRTX has expanded its range of supported LLMs (language learning models; i.e. GPT-4) to include new options like Gemma, the latest LLM from Google, and ChatGLM3-6B, a new open source option. The ability to choose your own LLM is invaluable, as different LLMs can perform the same tasks in completely different ways.
ChatRTX’s photo search and interaction tools have also seen a massive upgrade, with technology borrowed from OpenAI’s CLIP tool allowing the program to search images without the previously extensive and complex metadata labeling required by the application. Finally, ChatRTX users can now speak with their data, with added support for Whisper, an AI-powered speech recognition system that allows ChatRTX to understand verbal speech.
The technical demo leverages the power of Nvidia’s tensor cores and the TensorRT Library, so it requires an RTX GPU – and at the moment it doesn’t support the first-gen RTX 20 series either. So if you want to run the app yourself, you’ll need an Nvidia RTX 30 or 40 series GPU with at least 8GB of VRAM – and similarly, any professional Ampere or Ada GPU will also do the job. ‘affair.
The program will likely continue to expand its features ahead of a possible 1.0 release, including the expected return of its YouTube link cleaning capabilities that were removed recently, shortly after the 0.2.x emergency updates of fix a security vulnerability.
