ChatRTXknown previously as well as Chat with RTX on, now support more models to IT. Chatbot is NVIDIA, which was revealed in mid-February, which allows you to get the answer by using your own data as well as incoming calls. Privacy is guaranteed, because the information is not sent to the cloud.
ChatRTX support a model for other open-source
ChatRTX allows you to have a chatbot is a personalized to your computer. “The demo,” as defined by NVIDIA, uses, processes RAG (Take-Augmented Generation) for the optimization of the production of the LLM (Large Language Model), the technology TensorRT-LLM, developed by the company kaliforniane, and the power of the GPUS of the series RTX 40, and 30, a User must have a PC with one of the GPUS (at least 8 GB of storage), at least 16 GB of RAM, and Windows 11.
The performance of the local provides performance, security, and privacy. ChatRTX support the Mistral 7B of Mistral HE Called 2-Metës right now. NVIDIA has now added the Google's Gem, and ChatGLM3. The user can use the text, and documents the request (access), which was added to the pictures now, thanks to the knowledge of the conduct of the pattern CLIP (Contrastive Language-Image Pre-training) that have been developed by OpenAI.
In the end, by using the knowledge of the sound of a Whisper (OpenAI), is it possible to be sent to the incoming vocal in the chatbot. ChatRTX is free of charge and can be downloaded from the web site at NVIDIA. In addition to the requirements above, you should also be enough space on the disk (the ZIP archive of the voice 11,6 GB).
Discussion about this post