NVIDIA has introduced a new tool that simplifies the process of creating a personalized ChatGPT that runs locally on your PC. Known as Chat with RTX, this application leverages your computer’s hardware, specifically requiring a GPU from the NVIDIA brand.
Please follow us on Twitter and Facebook
Described by NVIDIA as a demo, Chat with RTX is not yet fully polished and may have some glitches. Nonetheless, it offers an intriguing alternative, potentially streamlining tasks such as searching for specific data in locally stored documents or generating summaries of YouTube videos.
The interface of the application is minimalist, and its operation is far from complicated. Once installed, users can choose between two large language models (LLMs) to power Chat with RTX: Mistral from the French startup of the same name, or LLaMA 2 from Meta.
Subsequently, users must select a folder containing the documents they wish to scan using NVIDIA’s software, and that’s it. From there on, they can ask questions to the chatbot, similar to using apps like ChatGPT, Gemini, or Copilot. The artificial intelligence will analyze the provided files and extract answers accordingly.
For instance, if Chat with RTX analyzes a folder containing files related to your upcoming vacation, you can ask it to indicate the date and time of your flight. NVIDIA’s chatbot will not only provide you with a text-based response but also link the documents from which it extracted the data for quick access.
Continuing with the same example, if you have a defined itinerary for your trip stored in a PDF, you can ask the chatbot to extract that information and summarize it within seconds. Moreover, Chat with RTX can also work with elements not stored locally, such as YouTube videos.
Creating Your Own ChatGPT with NVIDIA Chat with RTX
If you prefer extracting information from YouTube videos instead of local data, Chat with RTX allows you to do so effortlessly. Simply copy the link and paste it into NVIDIA’s application. The AI will process the content, enabling you to ask questions about it. Additionally, akin to documents, it can generate a summary in text format.
In the promotional announcement of this software, NVIDIA mentions that the application works with “selected videos.” While it doesn’t elaborate much further, it’s logical to expect varying results depending on the content. In simpler terms: don’t be surprised if the application frequently encounters issues when using this feature.
As mentioned earlier, Chat with RTX requires an NVIDIA GPU in your computer. The chatbot is compatible with all models from the RTX 30 and RTX 40 series with at least 8 GB of VRAM. Other requirements include Windows 11, at least 16 GB of RAM, and version 535.11 or higher of your graphics card drivers.
To truly leverage the app, you may need a fairly powerful PC. According to Tom Warren from The Verge, the app occupies about 40 GB of storage and utilizes a Python instance consuming around 3 GB of RAM when running. Lastly, it’s worth noting that if you use the app with local data, that information remains on your computer and does not get uploaded to the cloud.
Read Also:
ChatGPT Now Has Memory and Remembers Everything You Ask
OpenAI Seeks To Combat The Use Of ChatGpt For Political Purposes In The U.S. Elections