Learn How To Query Pdf using Langchain Open AI in 5 min

[@paulokafor1635]

I love this. Thanks Krish. you have really helped me learn machine learning. I started with your machine learning playlist, and now i am in deep learning. You are truly the best. Thank you, its now almost 1yr since i started this journey into data science, and i am so glad to have taken this step.

[@jasonbourn29]

Instead of open ai please...please use something we can access (free to learn)

[@somethingnew4363]

Use ollama

[@somethingnew4363]

Insted of openaiembeddings we can use ollamaembeddings it for free

[@alkebabish]

This stuff is incredible. It's got me back into coding again. Thanks for the clear video!

[@gautamkushwaha8724]

Love from Nepal. I am working on a project, and integrating pdf chatboat is very imoportant part of it. It was super helpful for me. And the way you have presented and explained the video was also quite interesting . Keep up the good work.

[@parthadhikari00]

This was short and crisp to understand, great video !

[@AnG-r3g]

Hi Krish. This is Stuff was Awesome. Thanks for the clear explanation. Could please make video on pdf query using Lamini, without OpenAI

[@Tech_Enthusiasts_Shubham]

Thanks a lot sir your videos are really helpful and also help me in enhancing my knowledge in ai domain hat's off to you for making such contribution for us❤

[@TRENDINGTECHSCHANNEL]

If we dont have openai plus subscription. Then what are the options avaialbe which are open source ?

[@ParthivShah]

Thank You Sir.

[@jimmybruneel4540]

Thanks very well explained

[@enough200]

this is the problem-- the previous video has api key explaination. but how to know which previous video.. googly fekta hai

[@chandrahaasvemula7251]

kzbin.info/www/bejne/qmPFhHSmorWti5Y Time-stamp: 1:25 import os os.environ["OPEN_API_KEY"] = " " put your openai key between the quotation marks

[@ankitchauhan925]

I love this tutorial. How we can preserve user question context for up to k consecutive questions?

[@onkarrajambatwar4930]

Is there data Privacy issues if we connect with OpenAI api? I know data coming from API is not used for training but I want to be sure there is another layer of security. Could you please tell us more about the victor store

[@bilalmalik1647]

You're a Bog Boss of LLMs😂❤

[@kushagrasingh7731]

Thank you, would love to see an example using pandas df and what are possibilities in terms of what all can we do with that using langchain

[@surajsharma9122]

Hi Krish, Thanks for sharing this informative video. Please make a video on data extraction from images/graphs,

[@bhaumiks.6543]

thank you for amazing content

[@subodhmhaske7829]

Best video bhaii

[@Sachin-xj1oq]

Suppose, I have to compare the responses of two PDFs. Example: There are two PDFs having an article about Cooking. What i want is to query our model "What is the process of making a tea?" Desired Response: Source 1: Method of making Tea and Source 2: Method of making tea. And after getting two separate responses, I am able to compare these results, efficiency and all.

[@kaleshashaik5959]

Do we have way to query from multiple PDF files ?

[@DivyaSinha-mp5si]

Hi Team, I am trying to replicate the google collab notebook in the description link of this video. I am facing an issue while using 'pdfminer' library in the 'open_filename' function. I also tried using 'pdfminer.six'. it gives the error as - "ImportError: cannot import name 'open_filename' from 'pdfminer.utils' (/usr/local/lib/python3.10/dist-packages/pdfminer/utils.py)". A resolution to this issue will be very helpful. Thank you.

[@iamsohan]

is it related RAG implementation?

[@sumitbanik8307]

Does Faiss work in all versions of langchain?

[@mohdnomaanalam1555]

Hi Krish, I'm stuck here, I've trained my llm model from pdf, following same steps you mentioned in the video. Now I want to use same response in my api which I'm getting from my model, please guide as I dont want to run the code to train llm model again and again.

[@rupeshverma3496]

Hi Krish - thanks for the video and could you please tell me how we can load it with html file content and do the relevant searches.

[@prazyraj1735]

I have this use-case where there are different types of documents. I can parse documents using document loaders using langchain. But, there are images also in these documents. I want to store them as metadata and if answer generated from a context chunk it show the image also. Please help.

[@KumR]

Good one... But got this warning LangChainDeprecationWarning: The class `OpenAIEmbeddings` was deprecated in LangChain 0.0.9 and will be removed in 0.2.0. An updated version of the class exists in the langchain-openai package and should be used instead. To use it run `pip install -U langchain-openai` and import as `from langchain_openai import OpenAIEmbeddings`. warn_deprecated(

[@ronit1822]

The pip install statement worked but for the import as, your way did not work but doing import as. from langchain.embeddings.openai import OpenAIEmbeddings worked.

[@ashutoshmaurya2714]

Hi krish can u please tell me how much accuracy of questions answer from that?

[@ranitbandyopadhyay]

excellent video

[@nehat786]

sir please do make a video on vector database where we can store our embeddings and than query from it using semantic search

[@alekhyakusam7476]

how to make the user query dynamic and read the pdffiles attached by the user

[@harshalmahajan5877]

Great content, Thank you for sharing it. I have a query - How does the model get trained if pdf content has been changed of the same pdf file? Will it maintain any versioning and give answers accordingly?

[@bruhm0ment767]

it doesn't get trained on the content, that's the thing, this method of 'RAG' was introduced as an alternative to fine tuning llm's on new data. Fine tuning requires loads of compute and data, and can't be done every time we want to add new data. And so, using RAG we just provide the data as context for the question answering task.

[@jas.sin83]

hi - can you provide an addon of streamlitting this?

[@A.M.8181]

"How can I make it so that I can continue to ask ChatGPT further questions, taking into account the response provided earlier? This is because the first answer to a question about a PDF is not always fully displayed, and I would like to say 'Continue'."

[@zunaidqureshi8521]

Hi, Krish great explanation, can you tell me how to query our text files

Is there a way to automatically generate those queries based on the PDF that we uploaded? Can somebody answer, please?

[@Jeganbaskaran]

Thanks for such a great video. Though we are talking about GenAI to many ppl, the first question am getting is about the security of the data. Could you please give more insight on that? Are the prompt questions saved on their side? Considering the real usecase we may ask some senstive informations about the oraganization, assets. In that case, user may not want to share their questions what they are asking. Kindly make a video Security on LLM

[@harshsinha9709]

I think indexes are created in the local for searching

[@AkhilKoduri_1]

Hey Krish, great information as always. Does this langchain can be used on confidential data as well, at this point?

[@gauravsharma-bt2mt]

Yes you can

[@oo_anonymous8672]

Your query n data will directly hit the openai data centres. For security of your data ..you can use through azure openai..

[@eashwarinfosys9856]

What if the PDF is an invoice document, like a billing doc or a Sales document, Tax receipts. Can we ask query related to the content in the invoice document.

[@Vmos710]

Not working for me

[@gouravsareen8126]

I'm working on csv data and created an index json file from the csv data and using this index json file i'm trying to get answers from gpt but it provides bad results, can you please make a turorial video on the same thing?

[@prashantgautam8983]

How can I download the document search and LangChain model trained in the above video for offline use?

[@lm10pranav]

sir can u also make a video on how to do this same with huggingface

[@prempanwar6195]

Hi I want to implement same things in the nodeJS, can do this? with same thing in it? could you please create any video on JavaScript ( node, react, next any of them), will be appreciated, thanks in advanced

[@gauravsolanki5612]

Hi Krish, If there are images in the PDF, can it read and answer questions related to that as well? Please help on this.

[@SRShankar-s5k]

its not working now AttributeError: module 'openai' has no attribute 'Completion'. Did you mean: 'completions'? :(

[@harshavardhan7097]

How to fix it bro

[@dnsorigamitutorials7017]

Run this command in your terminal: pip install openai==0.28. There is no attribute Completion in the newer versions

[@fasilmveloor]

Hi krish, thank you for the video, can you upload same thing as shown in the video with llamaIndex

[@mousumisinha2002]

Nice Video. Where can i see the file? 'budget_speech.pdf'?

[@pritishghatkar6148]

Hi Krish can you do the same thing as shown in the video using GPT4all ?

[@ProGamer-rs5xx]

Can you perform this using any hugging face transformers

[@raviyadav2552]

Is there any alternative where we can use any other model for free?

[@harshsinha9709]

Is there any way to chunk data based on paragraph/section so that we can create a pdf responder 🙄

[@akshatjaiswal6073]

what if we search for things that are not present in the pdf?

[@ShivShankarDutta1]

@Krish If I have scanned image pdf. Will it work?

[@ZohaibKhan-io6le]

Can I use the api key from openai which is free or do I need a paid subscription to use the key I am getting stuck with document search Getting error illegal b bearer Can you please help me in clearing my doubt

[@sudheervanapalli7262]

Will this works with only paid account on openAI?

[@bhaumiks.6543]

Will it work on scanned pdf?

[@revashukla4382]

sir the open api key is not working and it shows we need to pay for api key can we use gemini api key instead

[@work.aniketkumar]

did you find any solution for it.

[@aditimondal3995]

how to integrate it with streamlit?

[@PramodGunaware]

PyPDF2 can help with normal text but what about when it comes to mathematics books or any book that contains formulas or symbols what ways to extract these kind of data?

[@kavankomer3048]

You got the solution to your problem? Because images in pdf too can cause inaccuracies in the llm response

[@Coff03]

Hi..I am having error in the step embeddings = OpenAIEmbeddings() Pls help to rectify

[@arunjain5195]

same

[@atitbimali9224]

@@kuldeepkumarjsr just found Open AI doesn't provide any quotas for API calls to a free tier account.

[@varma1035you]

I am new to python. Which is the code editor that you are using here?

[@AadityaSPatil]

He's using Google colab

[@mohsinkhan-bw3cd]

Upload mlops tools videos

[@xoxomomo4540]

by 5 minutes he means watch at 2x

[@sauravchowdhury6383]

how to use this now in 2024, as davinci 3has been deprecated

[@emily_core.]

you can use gpt-3.5-turbo instead

[@benjoaoc.7723]

Hi Krish ! I'm looking to extract chapters from a PDF, including listing the chapters of the book, the page number of each chapter, and then dividing the book by chapters. I've tried using this method, but it hasn't given me the results I hoped for. Do you know of any other alternatives or libraries that could help? I'd be very grateful for your suggestions.

[@purvikabajaj8482]

Hi Krish, thankyou for the very amazing video and explaining it all very simply. Could you also please create a tutorial using CSV file?

[@vinaynalluri277]

If the pdf is not a row based how can we extract the info from the pdf, even if we try using easyOCR it will not be able extract info. what we have to do in such cases.

[@vijaypatil3734]

Do we have any sample code for chatting with pdf using Java or C#?

[@Kishangopal-x3c]

Hello sir can we create chatwithpdf without api key and offline

[@SudhanshuVegad]

Hi Krish. Awesome video. This really helped. Do you have any plans for creating a UI/webpage for querying and displaying results? Would love that!

[@learnpythoninds6023]

Hi krish. Can you do a session for summarization of pdf using langchain modules and prompt templates

[@avadheshkumarprajapati139]

Can any one tell me how to get Open AI api key?

[@taiworidwan194]

Thank you for sharing this informative video on using Langchain to query PDF documents. It's fascinating to see how language models like Langchain can be leveraged to extract valuable information from text documents, particularly PDFs.

[@AjithKumar-kn7dj]

Hi Krish, what about data embedded in images/graphs?

[@CelestiaSecurity]

It give me the details from outside of the pdf also. How can I restrict this with only PDF?

[@bevansmith3210]

You are amazing man

[@vishalgarg8423]

TypeError: multiple bases have instance lay-out conflict --- Please help in resolving this error I am facing this error in in langchain

[@tanajipatil2812]

this is really helpful, please come up with some videos regarding excel macro dealing with gen AI

[@Kishor-ai]

Using Langchain , I need to retrieve (Person name , like krish or krish naik) from my word document and I need to replace that word with any other name.

[@wiskerwhiskers]

I would like to know from which PDF the answer is coming. how can I do that?

[@kandamca2010]

@if you provide the spring boot ai realtime example, that will be great

[@rathodindranil]

Sir there are tools available which summaries entire pdf(any book in pdf format) can we do, something of same sort with langchain and openai, please do some projects on that🙏🏼

[@Developer367]

Can you please create an example of reading examples from a file into a dataframe, and ask question using prompt template and load the response from the prompt template into a dataframe, also do you recommend using Pandas dataframe agent?

[@sachinkr9764]

Hi Krish, will this model works on scanned pdf or on images?

[@madhusudhan4024]

but why it is take only small pdf file if it is large pdf file ??

[@sashankreddy2434]

We have 10,000 PDF file in multiple folders and its sub folders. Can we use lang chain to query from all the PDF files ?

[@zaheerbeg4810]

If I am wrong correct me, You have to convert your pdfs into vector data base and create your own model to achieve

[@dhanushnayak1674]

A video from digital sreeni, but the explanation of krish good so he is called as god of data science in India. But where krish stand out... Thanks You Sir for your teaching skills. You try max to make normal people understand. Much love god bless

[@sumainaisrar9170]

I have tabular data inside pdf files. This QnA does not comprehend tabular data inside pdfs. How can i solve this issue.

[@deveshmishra5245]

Hi, did you find any solution for this yet?

[@EkNidhi]

nice🙃🙃🙃🙃🙃

[@reinaldousadani1075]

What if the PDF have images? Do you have any ideas/advice on how to process that?

[@oo_anonymous8672]

You will need computer vision for that

[@swet_gokugod9382]

This is amazing man !!!

[@AbishekEkhool]

sir i have one doubt how can we convert math equations pdf document to txt

[@himanshu6002]

Did you figure it out?

[@akshatapalsule2940]

It was great learning! Thanks Krish

[@rishiraj2548]

Good day wishes

[@Carlisson1000]

Best KZbin Channel ever