Timestamps: 01:02 PDF demo (analysis of 1000-pages of annual reports) 06:01 Visual overview of the multiple pdf chatbot architecture 17:40 Code walkthrough pt.1 25:15 Pinecone dashboard 28:30 Code walkthrough pt.2

@johnpaulpayopay1074

Жыл бұрын

Instead of using an OpenAI API, can use local LLMs such as llama, alpaca, vicuna, koala

@MidoBroski

Жыл бұрын

The GitHub repository in the description does not include this multi pdf bot? Could we get the source code?

@chineseoutlet

Жыл бұрын

@@johnpaulpayopay1074 Agreed. I tried to use Vicuna7B to replace the OpenAI(my machine don't run 13B). It works OK, except it can't do query on the entire PDF. Like when I ask it to summarise the whole book, it just say "I don't know" trying to find ways to fix it. Any one here know how to fix it, please let me know.

@tangdavid3317

Жыл бұрын

Do you know how to fix this? - error Error: Cannot find module 'next/dist/server/future/route-modules/route-module.js'

This tutorial is legit the best. Custom Chatbots are going to be a huge biz op.

Thanks for this, your vids are some of the best on YT now for this type of thing as right now in this domain there is only two words for all the content out there - information overload!!! I am still going through your prev video on PDF ingestion and processing, thank you for sharing your invaluable information in a style that makes it understandable.

This is the best tutorial I've seen on embeddings search, the applications of this are endless, really excited to start building. Thank you so much for the work that you're doing :)

You are great at teaching the entire process! Please continue this series :) Thank you!

Just found this channel, 16k subs now, not for long. You are making great tutorials that I will say I should have made months ago. Keep making videos you can hit 2m subs in less than a year if you make content more relevant and useful to a wider audience and more entertaining

The excalidraw documents are invaluable to the presentation and really help visualize the flow of data between the pinecone database and langchain is used to make it all work. Thanks a lot for those diagrams, your efforts are not unrecognized. The possibilities of this are literally endless

Thanks for the detailed presentation and explanation of your concept! This is really exciting to learn - subscription is set, I'm looking forward to more videos. Thanks mate!

Woah! 😱These diagrams make it so much easier to understand concepts better than other KZread videos! Thanks for spending so much time and effort!

@Jingizz

Жыл бұрын

yes its so helpful to understand these concepts

This is amazing man. Thank you! It will take a while to fully understand how to implement this for myself but I appreciate the knowledge.

Cool stuff, It is just unbelievable how much depth and breadth all the computing field has become, Too stretched out and too spread, Too much subjects for to explore, Chating with books is a cool idea that could really help us access and organize scientific human knowledge in all fields.

As always, you're delivering awesome content !!

Thank you for this contribution to the common good. You are the man.

Thanks! Terrific walk-through, really appreciate this tutorial!

Thank you for sharing, please do more, your explanation is so easy to understand !

so interesting. as a non -techie trying to implement this is wild but fascinating

This is an amazing tutorial! More videos like this please.

Brother...! This is what I am talking about! My sellers are looking for this kind of capability for research as well as for searching our internal content management system.

Thank you so much. This is exactly what I need. To talk with a document, is kinda insane.

Thank you for your great videos, I really like them. One UX improvement comment: Try serving the PDFs as static content with your web application, then when you append #page=42 to the PDF's link, a user can go directly to that page in their browser. Very easy to implement and your users will save a copy and paste and 3 clicks.

@chatwithdata

Жыл бұрын

Nice idea, thanks.

best video i found for my personal project on the internet

Great videos! Im up for the course!

Thank you for the walkthrough!

Can you please show how the front end connects and how to build it? This is definitely the missing piece of the puzzle for me. Thanks for the videos. They're great!

Awesome stuff... Excellent video!

First off, very well explained, great detail in the video - thank you for taking the time to put this togeather. I was wondering, for use-cases where the documents you are working with should remain private... is there any way to create the embeddings with some locally run model, rather than sending the raw document text in chunks to OpenAI?

Thank you for this excellent contribution! It has really simplified this process for me greatly. One question: The first step (well, second step) in Phase II of your flowchart is to combine the Chat History with New Question. The next step is to submit this to an LLM to get a new standalone question. Why do you do add this step instead of immediately vectorizing the Chat History with the New Question. What does this "new standalone question" look like? If this is explained in this video, or another, could you please include the url? I'm kind of new to this. Thanks!

Hi there, what an awesome video. Will be trying this out this week, asap. I am wondering how long have you been coding, and more specifically how long have you been working with technology like in this video?

awesome tutorial, Thank you 🙂

I like how you diagram and explain things. Keep going. Any chance of having your template for diagrams for our personal use :)

Solid work!

Fantastic stuff my friend.

Great job. Thank you!

Hey, the tutorial is one of teh best of its king in this segment. Can please share the repo link to this multi pdf one. Its very useful for others to work on.

Hi, this is amazing! thanks for the contents Wondering if you have any idea on the scalability and and cost of having such a system?

@Chat with data , do you mind sharing the meta data version of the code? where you chunk and tag the documents by year and page number. I am really interested in what you showed in the demo. Impressive work!!!

Awesome tutorial

This is awesome man! Thanks for sharing!

This is golden 💛

thanks for this! great vid!

Thanks for this. It would be helpful to see your CustomQAChain Class. How do I tell Pinecone to only search through specific namespaces?

you are Legend brother

Will you be releasing the codes for this one? Loved the earlier videos.

Hi Mayo, is there a way to include paramaters in the conversationchain for maximum length, frequency penalty & presence penalty? Apologies if it's simple but new to this. With your instructions + video, i was able ingest my own PDF's. It's a great start & now i'm looking to optimise the outputs.

Hi, great tutorial! I was wondering if it were able to read the financial reports such as the net income, debt etc.. within the balance sheet or income statement?

Really nice work, thanks! How about loading files which are located remotely, like S3 bucket?

great tutorial, can we increase the response token limit?

This is massive!

awesome tutorial thanks for making this

Cool!! , Thank you!!!

Sorry im new to this channel looks amazing. How are you getting that chatbot? where can i find it? Amazing channel and information you are providing though

@chatwithdata thanks for sharing this demo. In my line of work we sometimes deal with sensitive and non-public data. What risk factors are associated with analyzing sensitive data using GPT apis?

Are you by any chance running a function that strips out the pdf: object [] from the metadata? For some reason when I try to upload my PDF, there is a pdf: object [] that is causing a Pinecone error. I noticed that you have pdf-parse as a dependency, but I can't quite see how this is working. Keep up the great work by the way!

Great video :) What do you think about using metadata filtering instead of separating data by namespace? By giving users the ability to select metadata filters, you can provide a personalized search experience with precise filters for two or more metadata types, without the need for feature extraction from text. Keep up the good work!

@jazzyj2899

Жыл бұрын

have you tried this? i have tried it in python. at least I've tried metadata filtering with the new SelfQueryRetriever, and it did not work. There is a Structured Query component using comparators that does not allow it to work without some extension of the class and modifcation. Let me know if interested in hashing this out together.

this is so impressive

Thank you bro... Your are a F*** genius!!!.. This is one of those things that will radically chage the world in a very short time

Amazing job! When database as source? Thanks

Great work! Now whenever people ask me how to do this I will just point at this video :D Cheers, thanks for shareing

this was great

This tutorial is mindblowing. How much does it cost to feed chatgpt all of the tokens from the context document though?

I don't think this is the right approach. It doesn't know enough context on the whole document. When you hit a prompt, he will match a chunk to your prompt, but what if the answer is contained in the context of let say 200 pages ? You could make something clever. You could aggregate chunk togethers asking the LLM to summarize and group them in "meta chunks", you could repeat the process until all years are contained into a single max limit tokens batch. Then, with the meta Data, you'll be able to perform a much more powerful search over the corpus, providing much more context to your LLM with different level of aggregation.

@user-jc3ys1yc2n

Жыл бұрын

Something like that already exists and is called a “recursive summarizer”. The hard part is figuring out how to actually perform search over the recursively summarized documents without invoking the GPT api an unreasonable amount of times except during preprocessing. I am currently doing research in the area of passage retrieval and would love to know your thoughts on this and possibly experiment with some of your ideas.

@Grahfx

Жыл бұрын

​@@user-jc3ys1yc2n Well, I had this idea while having my morning coffee, but I haven't conducted any deep research to support it. However, as you mentioned, there may be potential issues that arise. Have you considered using GPT4 to help solve this problem? Personally, I anticipate having to address this issue myself because I plan to summarize big HR data.

@JanBadertscher

Жыл бұрын

you lose more and more information by recursive summarization. better to use embeddings, so the LLM knows the whole context of your documents, no matter how many...

@Grahfx

Жыл бұрын

@@JanBadertscher No, of course, it doesn't have complete knowledge, that's the main issue. It only has partial knowledge based on a vector match. This is why it provides inadequate responses when the prompt is not very specific. It is incapable of extracting critical information that comes from reading the entire context. A human expert is currently better to do this.

@chatwithdata

Жыл бұрын

You can control the `k` returned source documents up to 1000 if you want to via Pinecone + gpt-4 context is 8k, so you can cover a lot of context across the doc.

Looking forward to LangChain Agents.

Your Excali diagrams are invaluable too. I’ve been trying to get into the habit of creating similar diagrams. Do you usually come up with them before you code, make them manually as you go, or put it together once you’re done?

@vinosamari

Жыл бұрын

I picked up the habit of doing it before and it helped tremendously. You might get sucked in for a while but it’ll help streamline the process

@chatwithdata

Жыл бұрын

Yeah usually before because it helps me think through how to solve the problem

@picklenickil

Жыл бұрын

Anything fancier than party tricks requires System design and engineering. Logic is almost always beautiful on paper

@gambaweb

Жыл бұрын

This is like having the architecture of your system m. Imo it is a must to have. Good job meanwhile

@PizzaLord

Жыл бұрын

@@chatwithdata what tool do you use for the diags?

bro which keyboard do you use, and thanks for this detailed video

Can you please use other Vector DB's too for the same example?

Thanks for the great video, I'm using it very well. When I use the code you posted on GitHub based on pinecone, is there any limit to the number of PDF files/total capacity/total number of sides, etc.? It worked fine when the total size of multiple PDF files was around 84MB, but when it went up to 128MB, I got a SocketError:other side closed error in npm run ingest, so adding a PDF file to an already existing vector store doesn't seem to work well. Do you think I need to adjust the timeout for undici to be a few times longer?

Thankyou so much

Hi, thank you for this great architecture, but I wanted to ask if there is a tokenization phase that happens here?

Cool stuff!May I ask how to use openai to generate new standalone questions from chat history and new question?😂Could you please be more specific?

Would building an app that allows doing all this through a simple UI (and with other features) infringe the ToS?

Hey! Cool video!! Have you added functionality for multiple pdfs to the GitHub repo? 🎉

hi, let try adobe pdf api , you can improve a lot your embeddings with that , the api auto detaching paragraph , title , table and image , and you can superpower your stuff , also the chunk in not more no sens text but is p or H1 ecc that have sense and you can alt give weight to data , for example h1 is more important of a end page little paragraph

I work in an academic research setting researching an extremely niche subject in material science and these kind of tools are what I have been waiting for. I can't wait for us to be able to upload the papers and books that contain all current available knowledge on the topic and be able to interact with that knowledge pool.

@PizzaLord

Жыл бұрын

I already built this tool. If you want to be a tester, let me know.

@bendrybrough4362

Жыл бұрын

@@PizzaLord I would like to.

@LimabeanStudios

Жыл бұрын

@@PizzaLord I would certainly like to hear more

@ahmeda8042

Жыл бұрын

@@PizzaLord I would like to.

@tylerd6962

Жыл бұрын

​@@PizzaLord how do I become a tester

This video and channel are awesome! I'm working on a version of this demo in python, would you mind adding an Apache or MIT license to the gpt4-pdf-chatbot-langchain repo so I can share without worrying about any weird copyright issues with github? (Accidentally posted this on the wrong video earlier...)

Does this only work for $TSLA? I would love to use this program for other companies. Thanks and wonderful tutorial 👍

Great video. But the problem with your solution is you have hit the OpenAI api at least 5 times and it would make it costly and not scalable. Other than that, a good project.

@chatwithdata

Жыл бұрын

Hmm where are you getting 5 from? The retrieval is from Pinecone, which is very cheap.

great video! can we see the source code for this? would love to see how extractYearFromQuestion is implemented

Will you be able to share the architectural diagram you provided in this video?

When will you release the code from this walk through (not just the "base template", but the actual code here that uses namespaces)? Also, do you know of any way to prioritize what source it uses? For example, if we have 3 PDFs for 2021, 2022, and 2023, and each declares a new color of the year, and I simply ask "what is the color of the year?" how can I ensure it gives me the most recent answer?

Can this be wrapped into a web app package for ease of use?

would be great if u could build this full tutorial in just python!

I wonder if something like this could be used to read/analyze/edit entire codebases with gpt-4.

Why not use Embeddings as the documents title so that instead using Gpt to do a search of the exact name space, it could do a semantic search of name spaces similar to the initial prompt, thus finding things relevant more effectively. Thoughts?

thank you! not clear how you select namespaces (pdfs). just by year in request? code in video not same as in github?

Great tutorial! Thanks for that. What about privacy concerns? Imagine we have a company with a lot of internal documents and wanted to use gpt to search through it? When training a gpt based on this as an external api it has access to docs, so I’m wandering what privacy department will say.

@chatwithdata

Жыл бұрын

Yes this is true, but large organizations will likely use the Azure OpenAI service which is tailored for compliance, privacy etc.

@MateuszTrojak

Жыл бұрын

@@chatwithdata So what about small startups? :)

@chatwithdata

Жыл бұрын

@@MateuszTrojak Could try open source alternatives like gpt-j but I haven't tested the quality.

Very nice video, can we use metadata filter instead of namespace , will it work? Also could you please share me your main.ts , fewshotprompt template..

What are your thoughts on allowing the model to compose additional questions after the first collection of document data is supplied from the search results? Similar to how the agent flow works in langchain?

@FelixLanzalaco

Жыл бұрын

you can setup gpt to keep iteratively asking questions to re-engineer the prompt. Search for ChaGTP the one prompt to rule them all.

Does anybody know if they let you retrain GPT4 with specific data ? I'm talking about something along the lines of transfer learning.

How is the amound of tokes used? embedding it and querying it?

Very nice information, can you please create a video for python based for the same? It will be helpful thanks in advance😍😍😍

I'm extremely interested in AI and Language Models, but I have a limited understanding of where to start. I was wondering if you'd be able to recommend any programs, courses, or tutorials for those in a similar position to me. Thank you, and what an interesting video this was!

@henry7434

Жыл бұрын

Maybe read the transformer, Bert and gpt papers first?

@1986xuan

Жыл бұрын

He has the course opened now. You will find the link the description below this video

@rajns8643

Жыл бұрын

If you have a basic understanding of multivariable calculus, probstats and linear algebra; see few videos on ANN, then RNN, then transformers and BERT, and then finally LLMs. Its a long way but this way you would be able to understand it more deeply and you would be able to understand the recent developments in LLMs more clearly.

Question: instead of PDF, could you point it to the database to do the same? How would one accomplish that? Thinking of a BI tool or an ERP. I think digitizing PDFs will be great for organizations that have technical machinery - example field crew sales - if they need to answer the technical question, it can be input (via voice) into GPT and reference company information. One would need to make sure it is secure (ie ChatGPT cannot store what it reads).

Hi, I found it pretty nice solution to this issue of filtering the database to search in specific documents when it is needed. I wondering if you tried to do it with Chroma? I am being pursuing something similar but I am using a selfqueryretrieval, bassically the chain receive the question and generate the retrieval to the vecstore adding some filters, in this case would be 'year':2020 for example. However it doesn't take the time to search with each filter, it just filter with an or statement. I like your solution but it is also true that you need to hardcode the names of the spaces. Maybe you can have another chain that create the name of the space automatically. Well, finally, I tried to see the code in the github, but looks like it is not updated.

Discord link invalid. Could you reshare? I am also looking for the latest diagram you were showing on the video. In git, I found one PDF scenario. Is it possible to re-upload that image? Thanks for the great content

How do you connect your openai api with the pinecone vectorstore

what scrapping all information on a website and feeding it into chatgpt or chatgpt can do it all by itself?

Thanks, but I am wondering is GPT-4 analysing a specific Chuck of a document or the whole document? Because in the diagramme is written specific document is sent to GPT-4 and we know that we have a token limitation

are the years hard wired in your code?

How will we be able to change the code to use weaviate instead of pinecone?

The GitHub repository in the description does not include this multi pdf bot? Could we get the source code?