Nvidia introduced Chat with RTX, a chatbot with generative artificial intelligence that runs on computers with the company’s graphics cards. In this respect reports Border.
Chat with RTX has the ability to retrieve information from YouTube videos by URL and retrieve text content. The service also supports searching by video transcript. The neural network can also process PDF files. According to the source, Nvidia’s chatbot outperforms Microsoft’s Copilot, which runs on Open AI’s GPT-4 language model, on this task.
One of the main disadvantages of a neural network is that it cannot remember context. Therefore, each new query will be independent and will not be able to reference data from the previous question or answer. To install Chat with RTX, you will need a web server and a Python instance using Mistral or Llama 2 models. The application takes up approximately 40 GB of space.
According to Verge reviews, Chat with RTX is still like a beta version with numerous bugs and limitations. In this regard, the neural network at this stage may attract the attention of developers and enthusiasts.
Previously recognizedThat Samsung will add live speech translation feature to its headphones using AI.