Please tell me how I can correctly transfer all my books and textbooks and documents in PDF format to a vector database while preserving the layout structure and equations? Maybe some of the people implemented this idea using Nougat: Neural Optical Understanding for Academic Documents (https://facebookresearch.github.io/nougat /)? If so, I ask you to say a few words about how you did it.
And let me ask you another question: how exactly does the ChatGPT Retrieval Plugin help you in the process of solving problems? Will it be possible to use it to extract information from your vector database during the ChatGPT dialog?
I am grateful in advance for the answers.
[link] [comments]