Build a RAG Based LLM App in 20 Minutes! | Full Langflow Tutorial
HTML-код
- Опубликовано: 4 июн 2024
- In this video, I'm going to show you how to create your own AI application that uses RAG (Retrieval Augmented Generation) without writing a single line of code - in just a couple of minutes, using Langflow.
If you want to land a developer job: techwithtim.net/dev
🎞 Video Resources 🎞
Langflow Docs/Install: docs.langflow.org/
Langflow Github: dtsx.io/3vR07qO
Build now with Astra DB: dtsx.io/3vZk6n2
OpenAI API Key: platform.openai.com/api-keys
Code/Flow File: github.com/techwithtim/Langfl...
⏳ Timestamps ⏳
00:00 | Overview
00:33 | Project Demo
02:14 | Setup/Installation
04:07 | Building a Basic Chatbot
09:27 | OpenAI Integration
12:33 | VectorStore Databases
15:00 | Adding RAG
21:35 | Testing The App
23:02 | Additional Features
Hashtags
#techwithtim
#nocodedevelopment
#llm
#RAG
Tim please please more more on AI, LLMs, LangChain, .... free APIs
yes, i want more and more on ai, machine learning
Please make one dedicated playlist Tim, which will contain RAG, AI agent, vector db-related concepts
Tim, I was trying to understand the RAG topic from the codes, but I was having difficulty understanding what the sequence was. It was a very clear and great video. Please don't deprive us of your knowledge. Thank you very much.
Very helpful, very clearly explained - thanks a ton!
Wow .This is awesome Tim.
Thanks for the knowledge tim😊
This is AWEsome, TY!
Honestly I've been waiting for an intuitive tutorial on RAG and it's really nice that you have posted this video.
Thanks @TechWithTim
my left menu is entirely different...mine looks like the one in the docs e.g "inputs" "outputs" I have neither
Great content - can you please clarify whether the input pdf can be also some unstructured text data or code ? If not,is there a way to to do that ?
Your explanation was excellent, thank you! Regarding the formatting of the PDF text, should it be structured as a series of questions and answers, or can I use a regular PDF ebook format and the system will be able to generate answers based on the book's content?
Thks Tim u r the best...
that was well presented..
Hi Tim, great video! I wonder if it is possible to export as python code or similar instead of json ?
Great!
how can we implement this in an app of our own, does langflow provide api for that?
I wonder, what's the real difference between Langflow and Flowise for instance? Are they just like the same thing, competing application for the same thing basically?
can we use this as a seprate microservice secondly integrate it with a hugging face llm, embeddings are cheap but q/a is expensive on opne ai
Can you make a video about creating a chat program with files that will be local without the Internet?
Thank you for your wonderful effort and excellent explanation in conveying the information.
For the langflow that we build, is there a way for get the source code of it?
I am using Hugging Face to run Langflow. I got to the first test run of the OpenAI connection. I have a funded account and valid secret key entered but get the following error after asking a question. Error building component error building vertex openAI: Openaimodelcomponentbuild() missing 1 required positional argument: 'openai_api_key' Any thoughts? Not much out in Stack or Git world.
i was literally watching a 1hr course of Langchain but the video is 11months ago then i saw your video😂
The Split Text LangChain component (@ 17:49) is experimental on Langflow's UI. And when used raises a ValueError: Error building vertex Split Text: 'str' object is not callable.
Which text splitter component did you guys use? Any help will be appreciated.
As of langflow 1.0.0a44 use the Language Recursive Text Splitter instead. I set the separator type as python and it worked for me.
Hi.
Can we add function calling also with RAG, using Langflow.
langflow is cool
Does anyone know if it is possible to run llama 3 with groq using this yet?
If this tool is free, then it’s a game changer
How can one make it as a stand alone product (webapp or native) with variable input pdf?
Is there a way to have the bot do a greeting in the beginning?
I keep getting the error "Could not install packages due to an OSError: [WinError 5] Access is denied:" when I try to install it - pretty certain admin privileges are there throughout. Any ideas why this is happening?
Hey Tim I didn't find the input option on the side bar
Imstall pre-release version
Hey Tim Just A Small Request Dude Hope You See This Comment, Can You Make A Cource On Tensorflow?
Please use open LLM flow for rag. Thanks.
Tim. Honestly your a beast bro!! What tips can you give me to learn python asap!!!???
How easy is this to scale??
It would be really helpful if the app can query a database instead of the ped doc and give results in natural language...
hey, i cannot see the input option on the side bar, pls assist me
Same. I'm seeing a much longer list of items on my sidebar than the one in the video, but nothing saying 'input'.
Install pre-release
Also what Tim is showing is the pre-release version which is quite different from the stable version (ie: you want have input block etc...). Simplest way to get this up and running on m1 mac is basically do " docker compose -f ./docker_example/pre.docker-compose.yml up` however you need to add "platform: linux/x86_64" to langflow service docker compose file ie:
docker_example git:(dev) ✗ cat pre.docker-compose.yml
version: "3.8"
services:
langflow:
platform: linux/x86_64
image: logspace/langflow:1.0-alpha
"How can one add this chatbot feature to the website?"
free APIs
i couldnt even able to install langflow have any one tried installing langflow its giving me so many errors
it did give too many too but venv solved the issue
@@DanielCupakTUR oh that's all? Virtual environment fix all those
@@DanielCupakTUR I've tried it in and out of venv and it still fails spectacularly.
@@shawn445 no it worked actually but u have to use the exact python version 3.10 it works only on this version
a-pah-chee 🤣
El chiste es programarla no usar un servicio
Sadly langchain is pretty mediocre.
What do you suggest then?