Nvidia unveils AI chatbot 'Chat with RTX' for local PCs – ReadWrite
Nvidia has released an early version of its Chat with RTX, which is a demo that lets users personalize a chatbot with their own content on Windows PCs.
The custom generative AI is now free to download, and it requires an RTX 30 or 40-series GPU with at least 8GB of VRAM. This means users will have the capability to upload their own documents in order to create summaries and receive relevant answers based on their personal data.
Calling all GeForce RTX users 📢
You can now create a personalized chatbot w/ the Chat with RTX tech demo đź’¬
Available to download now → https://t.co/5BmDMYdpkt #AIonRTX pic.twitter.com/GheslqdGQL
— NVIDIA GeForce (@NVIDIAGeForce) February 13, 2024
Chat with RTX can apparently let you search YouTube URLs and its transcripts for specific mentions or summarize an entire video.
Users will be able connect local files on a PC as a dataset to an open-source large language model like Mistral or Llama 2, enabling queries for relevant answers. According to Nvidia’s blog, Chat with RTX “uses retrieval-augmented generation (RAG), NVIDIA TensorRT-LLM software and NVIDIA RTX acceleration to bring generative AI capabilities to local, GeForce-powered Windows PCs.”
RAG is essentially an assistant that searches through the data, especially if the model set is particularly large. The tool supports various file formats, including .txt, .pdf, .doc/.docx and .xml.
The company, which became the fourth most valuable U.S. corporation this week, said by utilizing the power of local GeForce-equipped Windows PCs, users can enhance their experience and take advantage of generative AI with unparalleled speed and privacy.
“Rather than relying on cloud-based LLM services, Chat with RTX lets users process sensitive data on a local PC without the need to share it with a third party or have an internet connection,” it added.
Nvidia also invited developers to discover how RTX GPUs can potentially speed up large language models by consulting the TensorRT-LLM RAG developer reference project on GitHub.
In recent months, the company has experienced an extraordinary surge in its value. In December, Nvidia’s stock value tripled, outperforming every other company in the S&P 500. Its chips have proved crucial at a time when there is a global shortage.
However, there has been controversy over its sale of chips to Chinese military bodies and state-affiliated groups, despite a U.S. ban on the export of the commodity to China.
Featured image: Canva
Suswati Basu is a multilingual, award-winning editor and the founder of the intersectional literature channel, How To Be Books. She was shortlisted for the Guardian Mary Stott Prize and longlisted for the Guardian International Development Journalism Award. With 18 years of experience in the media industry, Suswati has held significant roles such as head of audience and deputy editor for NationalWorld news, digital editor for Channel 4 News and ITV News. She has also contributed to the Guardian and received training at the BBC As an audience, trends, and SEO specialist, she has participated in panel events alongside Google. Her career also includes a seven-year tenure at the leading AI company Dataminr, where she led the Europe desk and launched the company’s first employee resource group for disabilities. Before this, Suswati worked as a journalist in China for four years, investigating censorship and the Great Firewall, and acquired proficiency in several languages. In recent years, Suswati has been nominated for six awards, including the Independent Podcast Awards, International Women’s Podcast Awards, and the Anthem Awards for her literary social affairs show. Her areas of speciality span a wide range, including technology, Diversity, Equity, and Inclusion (DEI), social politics, mental health, and nonfiction books.
