The community member is trying to build a rag system over 1000 documents, each one roughly 100-200 pages, and in some cases exceeding 500 pages. They are required to use llama or mistral for data privacy, and they are trying to use a multi-document agent, but the llama index documentation is based on OpenAI. The community members in the comments suggest that the community member only needs to make a maximum of 2-3 line changes to switch from OpenAI to any other language model of their choice, and they provide a link to a document that may help with using llama locally.
Hi I am trying to build a rag system over 1000 documents each one is roughly 100-200 pages and some cases the number of pages exceeds 500. Additionally , I am required to use llama or mistral for data privacy. I am trying to use multi-document agent but llama index documentation is based on openai. How can I customize it to use llama or mistral ?
