ChatGPT v2.0: Chat with Websites, Search the Web, and Summarize Any Website | Chainlit Chatbot
Вставка
- Опубліковано 29 гру 2023
- #WebRagQuery is a powerful #chatbot, built with #OpenAI #GPT model in #chainlit user interface, that harnesses the power of GPT #agents, #functioncalling, and #RAG to offer an enhanced conversational experience. Here's how you can make the most of its diverse functionalities:
Normal ChatGPT Interaction:
Engage in natural conversations as you would with a regular ChatGPT app, experiencing seamless communication.
Chat with Websites using RAG:
Unleash the potential of the #RAG technique by sharing a website link with the chatbot. Initiate in-depth discussions and pose detailed questions about the content of the specified website.
Website Summarization:
Request the chatbot to provide a comprehensive summary of an entire website in a single interaction. Streamline your information retrieval process with this efficient summarization feature.
Web Search with DocDocGo Integration:
Access the vast knowledge of the web by leveraging the chatbot's connection to the DocDocGo search engine. Simply input your query, and watch as the chatbot fetches relevant links and information from the web in response.
00:00:21 Related works (RAG-GPT and Web-GPT)
00:03:39 Chatbot Demo
00:10:20 Project schema explanation
00:18:18 Project coding structure
00:20:00 Explaining the code part 1
00:34:23 Socket usage error solution
00:36:29 Chatbot memory explanation
00:46:15 Explaining the code part 2
01:10:38 Wrap up
🚀 GitHub Repository:
LLM-Zero-to-Hundred Project: github.com/Farzad-R/LLM-Zero-...
WebRagQuery project: github.com/Farzad-R/LLM-Zero-...
📚 Main Libraries:
#OpenAI: platform.openai.com/docs/models
#chainlit: www.gradio.app/docs/interface
#Langchain: python.langchain.com/docs/get...
#Chroma: docs.trychroma.com/getting-st...
#duckduckgo-search: pypi.org/project/duckduckgo-s...
📺 Introduction to Text Embedding:
Watch the Video: • RAG explained: A Step-...
📺 Introduction to Function Calling:
Watch the Video: • Video
📬 My contact info:
/ farzad-roozitalab
farzad-r.github.io/
#OpenAI #GPT #chatbot #chatgpt #functioncalling #agent #rag #EnhancedChatbot #WebInteraction #InformationRetrieval #AIInnovation - Наука та технологія
Wow, this is a golden gem iof a channel!! I appreciate so much you sharing all of these projects, and more importantly diving in and walking us through the logic and analysis A++ Channne;! all VALUE no FLUFF! YOUR THE MAN!!!. At this rate and of this quality you'll be able to ask for a sale on your channel from your subscribers with huge rewards and up sides. Id buy from you for SURE!
Thanks Jason! that means alot. I am glad that you liked this video. This is my favorite chatbot so far but received very little attention.
But in case you liked this one, next project that I am working on would be very interesting for you I guess. That would be the biggest and most advanced chatbot that I have planned to upload on youtube and also haven't seen a chatbot close to it yet anywhere personally. It will use 5 different models to provide an any-to-any user experience. You can interact with it using text, voice and by sending an image, and it will respond back by text, voice, and generating an image if necessary as well. + it will include all the abilities of the chatbots that I have uploaded so far (searching the web and chatting with websites, RAG, etc.). That would be the last and the biggest chatbot for this series.
And with regard to the sales, my main goal has been to give back to the community from which I learned a lot. I have never thought about sales TBH. If I can make it to monetization or to earn from the Ads, that would be nice. But thanks for the + energy and your kind words!
Thank you for the clear explanation. Keep up the great content, subscribed!
Appreciate the kind words Emre! 😉
Good stuff!.👍
Great stuff... Please make a real time chatbot which can chat with mysql database which have multiple tables like 10-15.
I was just thinking about it today. I am now working on two other projects that will soon be uploaded on UA-cam. But will put it on the list to build it as well. Thanks for the suggestion!
@@airoundtable does it have functionality to chat with pdf's........can i click on attach button load pdf and ask que on it?
@@SonaliBhanudasMali This chatbot is specifically designed for interacting with the web and duckduckgo search engine. So, you cannot perform RAG with PDFs with it. However, you can easily integrate that capability in this chatbot. I have multiple projects with full RAG capabilities, please feel free to check them out. You can also combine them together if that is your objective. Here are the videos in which you can find Full RAG capabilities including chat with uploaded documents:
ua-cam.com/video/1FERFfut4Uw/v-deo.htmlsi=_j-V_YOBqen71BP2
ua-cam.com/video/AAvqscJPXIY/v-deo.html
ua-cam.com/video/6dyz2M_UWLw/v-deo.htmlsi=hXCv2p0FHwp_qqQX
ua-cam.com/video/nze2ZFj7FCk/v-deo.htmlsi=onPlVgJVffp0WyZe
Cool!
I have a quick question. I was able to get this WebRagQuery to work but when I setup RAG-GPT in the same fashion using Azure I received the following error message. I was unable to track down where "deployment_id" was declared as well. Any help is much appreciated! Thanks!
Error: openai.error.InvalidRequestError: Must provide an 'engine' or 'deployment_id' parameter to create a
Hey, Mike. Great to hear it! I just didn't understand from your message whether you are using azure openai or openai itself. All the issues are rising from the fact that there are more than one ways to generate response from GPT models.
1. If you are using azure open ai, when you deploy a model on your azure openai platform, you assign it a deployment name. You can see it if you go to your oepanai studio and check the deployed models. So, for generating the response, you have to pass the name of the model (e.g: gpt-35-16k) to "engine" argument. And you have to pass the deployment name of that exact model ("e.g: my-azure-gpt-35) to "deployment_id" argument.
2. If you are using openai directly, again you have to use their provided code for generating the response from GPT.
Ah, ok. I see. Thank you very much for the clarification. I will give that a try. Thanks again!@@airoundtable