PDF Table extraction for RAG

[jais001]

Hi,

I am using R2R version 0.2.84. The extraction method I am using is unstructured "by_title". I have pdf containing lot of tables, while using this chunking config and performing RAG, i have noticed that the accuracy of RAG querying is so low. is there a way to increase the accuracy, so i can perform querying from tables in much more accurately?

Thanks for the help in advance.

[shreyaspimpalgaonkar]

Hey, you can set the pdf extraction to hi-res instead of fast or auto. Also, you can use unstructured's API for even higher quality outputs.

[jais001]

@shreyaspimpalgaonkar But the hi-res only works with the latest version of R2R, right? it is not implemented in R2R version "0.2.84" right?