Nvidia, demo application Chat with RTXReleased an early version of. The application allows you to run a personal artificial intelligence chatbot on your computer. You can feed Chat with RTX with YouTube videos as well as your own documents. This way, you can make summaries and get relevant answers based on your own data.
The entire process runs locally on a personal computer (PC). At this point you only need a GeForce RTX 30 or RTX 40 series GPU with at least 8GB of VRAM.
Highlights of Chat with RTX
Chat with RTX artificial intelligence initiative within default settings MistralIt uses ‘s open source model. In addition, the vehicle Meta’s Llama 2It also supports other text-based models, including . According to the information shared by Nvidia; Downloading all the necessary files takes up 50GB to 100GB of storage space, depending on the models chosen.
Chat with RTX works with text, PDF, .doc, .docx and .xml formats. When the application is directed to a folder containing supported files, the files are loaded into the model’s fine-tuning dataset. Additionally, Chat with RTX can take a link to a YouTube playlist and upload transcriptions of the videos in the playlist. However, you can query the contents of the videos in question on any model you choose.
Despite all these features, the vehicle also has various limitations. Chat with RTX cannot remember the context. In this sense, the application does not take into account previous questions when answering follow-up questions.
Additionally, the relevance of the application’s answers can be affected by factors such as the question sentence, the performance of the selected model, and the size of the fine-tuning dataset. According to the information shared by Nvidia; response quality can often improve with larger data sets. Chat with RTX can then direct users to more content on a particular topic.
When viewed this way, we can say that Chat with RTX is far from competing with the giants in the market for now. Still, the ability for the tool to run locally on the computer offers a significant value proposition. Because the data processed by offline models does not leave the device on which they work. These models, which are also more cost-effective than cloud-hosted models, also stand out with lower latency. We will all watch and see how Chat with RTX will be positioned in the coming period.
Source link: https://webrazzi.com/2024/02/14/nvidia-chat-with-rtx/