"How to 10x chatbot UX? 🤖 🖼️ " - Add Image Responses to GPT knowledge retrieval apps

Ғылым және технология

A step by step tutorial of how to enhance your chat with PDF/website app user experience by enabling it to respond with not just text, but also images 🔥🚀
🔗 Links
- Follow me on twitter: / jasonzhou1993
- Join my AI email list: www.ai-jason.com/
- AI Researcher Github: github.com/JayZeeDesign/knowl...
- My discord: / discord
⏱️ Timestamps
0:00 Intro
0:43 Why GPT chatbots don't return imgs
1:23 Case study overview
2:20 Step 1: Scrape HTML
4:03 Step 2: Convert HTML to Markdown
6:22 Step 3: Create vector index with llama index
7:10 Step4: Retrieval Augmented Generation (RAG)
8:43 Handle PDF docs
👋🏻 About Me
My name is Jason Zhou, a product designer who shares interesting AI experiments & products. Email me if you need help building AI apps! ask@ai-jason.com
#gpt #chatgpt #ai #artificialintelligence #tutorial #stepbystep #openai #llm #langchain #largelanguagemodels #largelanguagemodel #chatwithpdf #autogpt #chatgpt4 #gpt4 #gpt3 #aiautomation #aiautomation #aiagents #nocode

Пікірлер: 59

  • @jaysonp9426
    @jaysonp942610 ай бұрын

    Dude, your content is fire. You keep finding use cases and angles no one else on KZread is covering

  • @zidviziouz

    @zidviziouz

    10 ай бұрын

    Couldn't have said it better myself!

  • @user-vk5fj5mj8s

    @user-vk5fj5mj8s

    10 ай бұрын

    Yes, I agree. Unique content. I'm looking forward to new video posts. Jason is cool 😎

  • @PhoebusG

    @PhoebusG

    10 ай бұрын

    Agreed, gained a loyal fan here.. keep up the good work ^_^

  • @MarkRambotech

    @MarkRambotech

    10 ай бұрын

    agree

  • @AhmedIbrahim555
    @AhmedIbrahim55510 ай бұрын

    9:26 Yes please, 💯%open source, that what I love your channel for.

  • @mrdbourke
    @mrdbourke10 ай бұрын

    Jason your channel is outstanding for interesting and helpful use cases of LLMs, I've been experimenting with your techniques and loving them. Thank you and please keep it up!

  • @kchapadia
    @kchapadia10 ай бұрын

    Incorporating images in responses definitely improves the UX. Would love to try this out with PDFs, an open source library to convert PDF to Markdown would be great!

  • @fire17102
    @fire1710210 ай бұрын

    Foss Pdf2Markdown sounds incredible! Please explore the idea further ❤🙏 When you do, please remember to design it in a way where we can easily jump back to (and hopefully highlight) the original source at the right place. Either on the Markdown (easier) or to the PDF (more complete) Thanks a lot for everything, You run one of the coolest channels for sure

  • @MeanGeneHacks

    @MeanGeneHacks

    8 ай бұрын

    ❤ 100% Open PDF2 markdown is needed

  • @jameslibby5215
    @jameslibby521510 ай бұрын

    I would love an open source pdf to MD library. I'm working on creating a "chat with your documents" program for my company so we can use our confidential data and that would be so useful. I'm a young CE, so I'm still so ignorant on so many things. Subscribed. Great content.

  • @jasonfinance
    @jasonfinance10 ай бұрын

    This is soo good, having rich media as part of response really improve the UX of the GPT apps!

  • @half_way_expert
    @half_way_expert10 ай бұрын

    Another great tutorial Jason! Keep it up and keep generating these new ideas!

  • @garmeeh
    @garmeeh10 ай бұрын

    Keep up the good work! Content is always on point 🚀

  • @rickylafleur4378
    @rickylafleur437810 ай бұрын

    I subscribed to couple of newsletters about ai but your videos are just simoly better than all of them

  • @onangarodney7746
    @onangarodney774610 ай бұрын

    I'd love to see the Pdf to markdown. You are doing a great job 👍🏾

  • @adrianwhite100
    @adrianwhite10010 ай бұрын

    Awesome work Jason. Straight up smarts period.

  • @pratiksingh3432
    @pratiksingh34329 ай бұрын

    Straight up, best content. Keep it up and we sure do need an opensource pdf to markdown converter.

  • @asithakoralage628
    @asithakoralage62810 ай бұрын

    Fantastic video Jason, thanks a lot, I’m learning a lot from your channel.. thanks a lot

  • @user-lv7ft2zu7k
    @user-lv7ft2zu7k9 ай бұрын

    Excellent job of explaining the material. You make learning this stuff enjoyable. Keep up the good work and hell yeah, deep dive. Thanks

  • @user-vk5fj5mj8s
    @user-vk5fj5mj8s10 ай бұрын

    Good job Jason. Thx 🙌🏾

  • @roshghadamian733
    @roshghadamian73310 ай бұрын

    This is excellent! Loving your content, would love to have an open source pdf to markdown package!

  • @ajitdmonte
    @ajitdmonte10 ай бұрын

    Very nice content😊

  • @Afzal3000
    @Afzal300010 ай бұрын

    Thank you so much

  • @user-iv9om2ld2l
    @user-iv9om2ld2l10 ай бұрын

    9:28 great job! I’d love to try an open source pdf to markdown library.

  • @seanxjin
    @seanxjin9 ай бұрын

    That’s some real value!

  • @RussellLundstrom
    @RussellLundstrom9 ай бұрын

    super impressive Jason, do you do any custom dev?

  • @micbab-vg2mu
    @micbab-vg2mu10 ай бұрын

    thank you for the video

  • @A28480
    @A284808 ай бұрын

    Pdf to markdown would be fantastic!!

  • @stepkurniawan
    @stepkurniawan9 ай бұрын

    awesome vid man! btw, an idea, what about using AI to update your notes automatically? :D

  • @OryginTech
    @OryginTech10 ай бұрын

    Bro I’ll make a shrine of you in my house if you make an open source version of that converter 😭

  • @briangrover1625
    @briangrover162510 ай бұрын

    🔥

  • @bahamu
    @bahamu10 ай бұрын

    I would love PDF to markdown!

  • @albertalbert5785
    @albertalbert578510 ай бұрын

    PLS MAKE A PDF TO MARKDOWN LIBRARY

  • @sel3460
    @sel34609 ай бұрын

    Jason, brilliant work, please continue on pdf2markdown and come out UX with steamlit or gradio

  • @chivesltd
    @chivesltd10 ай бұрын

    I'd love to see the Pdf to markdown!

  • @musumo1908
    @musumo190810 ай бұрын

    This rocks. Yes to an opensource PDF markdown. How do you suggest creating an API for this and any of the coded langchains ? I want to test dropping the API into a bot builder…the builder probably won’t support img response….will also drop into an appbuilder. Thx

  • @sanchaythalnerkar9736
    @sanchaythalnerkar97369 ай бұрын

    How is LLama Index different from using the traditional emebddings and vector store?

  • @robinwang6399
    @robinwang639910 ай бұрын

    I am currently working on a tool to extract tables from pdf as data frames and images as image references from pdf. I was wondering how I can display them, and happened on your videos. This is very helpful. I am currently having trouble extracting PDFs with images of mathematical formulas and expressions, OCRs from what I can see tends to give nonsense. Due to how messy PDFs are, extracting image and table contexts and dynamically inserting them into documents that doesn’t have these data becomes a pain in the backside. (There was this PDF which had 5 to 6 images in a cluster, and displayed as a single image with texts and math around them, there was only one description for the cluster, my tool only managed to extract the first image and group it with the description.) I am thinking of finding AI models that generated texts from image, to dynamically generate the contexts alongside using OCR, but that can become really slow. If anyone have any ideas, please help.

  • @gawrysiak100
    @gawrysiak1004 ай бұрын

    thx

  • @ithaca2076
    @ithaca207610 ай бұрын

    please we need a PDF to MD library that would be incredible

  • @betun130
    @betun1309 ай бұрын

    I would be very interested in a pdf to markdown library version!

  • @yixuanyang8516
    @yixuanyang85168 ай бұрын

    a quick question: does gpt 3.5 understand and analyze the img/video that's embedded in the html? because if so, then it's kinda doing what llava does

  • @yashsrivastava677
    @yashsrivastava67710 ай бұрын

    please create an open source version of pdf to markdown library.

  • @jacobisknight
    @jacobisknight8 ай бұрын

    i want that, would also be interested in building that for open source.

  • @dharmendra_kumar_yadav
    @dharmendra_kumar_yadav10 ай бұрын

    Bro thanks I am also finding this can you make an ai chatbot that can give image and give output and also talk using model chatbot 😢❤❤❤

  • @srh1034
    @srh10344 ай бұрын

    @AI Jason Amazing video! Can you please teach how a CPU-based chatbot can be built without needing OPENAI and any other API key in case a user like me want to host the solution and not transfer sensitive data to cloud? And also minimize the cost of having a local CPU-based chatbot?

  • @hocklintai3391
    @hocklintai33917 ай бұрын

    I tried it it, but the rendered markdown's images are not perfect. I applied it on some other sites, and the images may failed to appear in some cases.

  • @boypotus
    @boypotus10 ай бұрын

    Yes many regurgitated tutorial | lacking in richness of experience agreed

  • @lilspyguy
    @lilspyguy8 ай бұрын

    How would you add chat history to this?

  • @WN1917
    @WN191710 ай бұрын

    CAN you make a video where you train llm with pdf document containing exam questions and answers. Then the user could use that llm do generate questions and answer to question. Kind of an ai tutoring platform

  • @xavierf2229
    @xavierf22299 ай бұрын

    Pdf to mark down library please bro

  • @AhmedIbrahim555
    @AhmedIbrahim55510 ай бұрын

    I,m just software engineer, what the exact necessary roadmap should I take to do what you doing.please😊

  • @AIJasonZ

    @AIJasonZ

    10 ай бұрын

    Thanks mate! Id start with the prompt engineer :)

  • @AhmedIbrahim555

    @AhmedIbrahim555

    10 ай бұрын

    @@AIJasonZ I appreciate your response,thx 😊

  • @python_handler
    @python_handler10 ай бұрын

    create one video on research GPT using open source model ...not open AI. thanks in Advance..

  • @tijendersingh5363
    @tijendersingh536310 ай бұрын

    Nice.. can we have the code file

  • @emmanueldomains2829
    @emmanueldomains282910 ай бұрын

    The most impactfull youtube channel I have come across. @AIJasonZ I sent you an email

Келесі