Chunk large complex PDFs to summarize using LLM

In this video, I talk about a technique to do context aware chunking of large PDFs and then summarize the content using map-reduce framework(implemented through Langchain)
References:
arxiv.org/pdf/2307.03172.pdf
developer.adobe.com/document-...
python.langchain.com/docs/use...
smith.langchain.com/hub/
code:github.com/rajib76/langchain_...

[@fabsync]

oh man.. thanks for your videos! They are precious gold! I love the way you think and teach!

[@user-on1qo8un2t]

Man, very great explaination. I m constantly visiting your channel for great new tutorials @@@:)

[@ianmatthews8342]

Excellent, nice idea and very well explained! Thanks!

[@virunew]

nicely explained!

[@aarshmehtani5468]

Thanks Sir. Actually I am working on some project based on it and find difficult for me to find materials to understand the concept practically. After watching this video, I understand and implement it successfully and step ahead.
At last Thanks Sir for this video.

[@user-ll6mn8ky8d]

you'll be famous soon

[@timtensor6994]

Maybe for image, how does GPT4 multimodal models work ?

[@bilalzahoor5608]

Hi Rajib, Thanks for making this video. It has been really helpful as I try to build a RAG system for a B2B use case. However, I did try setting up the Adobe API but I must say it's not too easy as I am getting stuck at various steps. I am not able to get 201 response code. Can you please share the steps you followed to setup the API? Regards, Bilal

[@PrabakaranSPpraba984]

Hi , Could you please share your LinkedIn profile ? I am doing the same PoC , I need some clarification.

[@elephant742]

Hi Rajib,
Really insightful video. Especially the Extract API for the context-aware extraction of text from PDF.
Are you aware of any open-source alternatives for the Extract API?
Regards,
Dev

[@rajibdeb4059]

:) I was also looking for one. Tabula, Camelot did not work for me. Looks like the enterprise grade solutions come with a price.

[@MadhanAnbalagan-ff5qt]

Closest open source I found working is unstructuredIO

[@rajibdeb4059]

that also did not work for me, in fact i have shared the results with unstructured
@@MadhanAnbalagan-ff5qt

[@user-ll6mn8ky8d]

if you had to pick an open source alternative for this, what would you chose? ​@@rajibdeb4059

[@loknathbasak1533]

@@MadhanAnbalagan-ff5qt Did you find a okaish opens source one?

[@vikasrajpurohit8730]

Thank you, can you give your LinkedIn handle