Artificial intelligence has managed to obtain a position among the most striking technologies of recent times, not only because of what can contribute, but also because of the ease that users can have to access it. But in many cases, it is possible that there are people who do not want to pay a subscription model to be able to access, for example, to a chatbot As is the case of chatgpt, they must therefore look for other alternatives for this, and one of them would be Chatrtx of Nvidiawhich makes it possible to implement a chatbot on a PC which has an RTX graph of the brand and to execute it completely locally.
There are a lot of features that we can find on the Internet that allow us to take advantage of new technologies to access various resources in a much simpler way. THE chatbot For example, it facilitates users the possibility of accessing the summary information they can find on the Internet, as well as generating various texts via a sentence or a question. And although we think that it will generally always be necessary to access an online resource for this, we can also use it locally thanks to applications that are like Chatrtx.
A chatbot directly on your PC, this is how Chatrtx works
There are many applications that allow us to have the latest technologies in our hands, although it is often necessary to have specific equipment that allows them to use. In the case of Chatrtx This is something similar, because its name indicates that this program uses Nvidia graphics to offer the user the possibility of creating a chatbot in his Pc Using various AI models such as Mistral 7B int4, chatglm 3 6b int4 clip.
Although, obviously, it offers the characteristics that we can find in any chatbot has a series of problems with regard to the most advanced models that we can use on the Internet such as Chatgpt. Created locally, it uses the benchmarks that have been installed with it, this implies that the information really collects it from the documents that we have on the computer itself. In addition to that when you use our PC equipment if we have a chart RTX It will take a long time to generate the content while forcing the use of this 100% component while generating the answer.
On the other hand, he can also present failures in certain responses, such as being stuck to repeat the same sentence for several paragraphs, but it is always a fairly curious application that allows us to know a little how these Models work, because as well as personalize them in a certain part.
The installation and use process is quite simple:
- We download the application from the NVIDIA official page
- We decompress the Chatwithrtx folder where we want to store
- We install the application by clicking on “Configuration”
Once it has been installed with the parameters we want, we can use it, if we click on the option indicated by “IA model”, we can modify between the three models that we mentioned below Above, while clicking on “data”, this will allow us to change the file from which we want to collect the information. This offers us the possibility of saving various documents .txt, .pdf and .docx in a folder to use them as a reference for Chatbot itself, that is to say that we could, for example, download the sheet of a motherboard and integrate it into the file to be asked for the model on the different aspects of this component.
