Catch remarkable information from any Object

[granite knoll]

Hey Guys!
I am working on AI company and I am working as a NLP Developer.
I am trying to catch all remarkable information (ex Stock News, workforce) from any objects and it can be text, pdf file, voice, audio and so on.
Can I overcome this challenge using OpenAI or LangChain?

[frank radish]

langchain does have tools available to help you parse rough data objects (ie PDFs). I assume that if you are pulling data from the internet (web scraping), you have the option to already format and store your data in structured objects.

Another thing that you can do with langchainn is create retrieval augmented generations with some models. You can build a database with your objects and have langchain perform a nearest neighbor search on that data via embeddings

Good news is that langchain does have integration with OpenAI API so you can use their models as well

[granite knoll]

Thank you
COuld you provide me the reference - github or tutorial?

[frank radish]

LLaMA 2 70B to rebuild GPT Banker YouTube video: https://www.youtube.com/watch?v=SedGB8m2XLM&ab_channel=NicholasRenotte

LLaMA 2 70B to rebuild GPT Banker GitHub: https://github.com/nicknochnack/Llama2RAG