Step by step no-code RAG application using Langflow.

Рет қаралды 14,582

2 ай бұрын

Source code of this example:
github.com/svpino/llm/tree/ma...
Lanflow's GitHub Repository: bit.ly/3JcNXeC
Astra Vectorstore database: dtsx.io/4aw3x17
I teach a live, interactive program that'll help you build production-ready Machine Learning systems from the ground up. Check it out here:
www.ml.school
To keep up with my content:
• Twitter/X: / svpino
• LinkedIn: / svpino

Пікірлер: 58

@dragonmax2000 2 ай бұрын

What I liked about this is that you start with intent and end with a result, while meticulously going through specifics, even if it is high level.

@underfitted 2 ай бұрын

That’s the goal!

@dragonmax2000 2 ай бұрын

I am a software engineer with looooong experience, but I am super confused about AI and inner workings. I might just start this was to see how things operate at the top and turn this into code on the other side. Make more videos about langflow, very nice content. Thank you!

@coenraadloubser5768 2 ай бұрын

$10 per step...? This is what makes all the incoming finetunes of Lllama3 70B a game changer... that price will drop to

@DataStaxDon 2 ай бұрын

TY for the thoroughness in your explanation!

@user-ks6uq1eu2w 2 ай бұрын

From learning one day to using what I learnt to quickly develop a POC is super cool. Thanks Santiago!

@laugustou 2 ай бұрын

Esto es oro Santiago, muchas gracias!

@AhmedRehan-fz1eq Ай бұрын

Perfect , this really opens my mind about the possibilites we can have . keep it up , these type of videos really help

@MeinDeutschkurs 2 ай бұрын

Awesome! Thx for sharing Langflow. The visual approach is my wqy to go. Exactly what I was looking for. And boom, it appeared in my timeline. 😎

@ThienHaFlash 24 күн бұрын

Very good and clear explanation! Thanks a lot for sharing the valuable information!

@stanTrX 2 ай бұрын

Wow man. This is awesome!! El fantastico. Cant wait to have more tutorials, with fully local options this time. 🎉

@AJICAR 2 ай бұрын

This is a truly excellent tutorial. It's the first time I've worked with this framework, and the video covers every detail with perfect clarity. I made some variations, using Ollama with llama3 and Chroma as a vector store, and everything went smoothly thanks to your clear explanations. Fantastic work!

@underfitted 2 ай бұрын

Awesome!

@seanb9949 2 ай бұрын

And yet another great post Santiago. This is awesome!

@user-xw7or3si5b 2 ай бұрын

Exactly what I was looking for! thank you!

@mauricioferreira8344 2 ай бұрын

Great content Santiago! Well done and structured! Thanks!

@joaquinduran7407 Ай бұрын

Really cool, thanks for sharing!

@lusam8294 Ай бұрын

Great video, that's exactly what I'm looking for. Thank you so much for sharing！

@Azam.Rajabov 2 ай бұрын

Thank you so much for the great video.

@user-hm6pp4rd3z 2 ай бұрын

This is awesome, great content, perfect explanations, so helpful, 1000 thanks Santiago for sharing! 🙏

@germainrodrigue367 2 ай бұрын

Amazing 🤯🤯🤯, Thank you Santiago

@aousabdo7266 2 ай бұрын

Always great contents! Really clear and concise demo of langchain. Maybe you can make a video of how to allow the user to dynamically chose a website to ask questions about

@tarangsuri8932 Ай бұрын

Great tutorial...

@bAvt_xyz 2 ай бұрын

Great tutorial. I'd LOVE to see someone (anyone!) demo how to use OpenAI custom agents via API...not just the vanilla core model(s) API.

@TarabUTK Ай бұрын

Awesome session .... I coded along and LangFlow is pretty cool.

@jkhhahahhdkakkdh 3 күн бұрын

You did?!

@marcelomigueletto 2 ай бұрын

Really good!

@malartbecomes236 2 ай бұрын

Holy crap this project looks SO MUCH BETTER than Flowise, from top to bottom.

@mandelafoggie9359 Ай бұрын

Nice, thanks

@TPH310 2 ай бұрын

Thanks man

@morphos2 2 ай бұрын

Excelent!!

@topertyco 2 ай бұрын

Hi! Thanks for the tutorial. How do you recommend to calculate costs to run the application in production?

@TooyAshy-100 2 ай бұрын

Thank You Santiago Thank you for sharing your insights on "Step by step no-code RAG application using Langflow." Question: Regarding your use of chunking by 1000 and embedding with a size of 1536 for OpenAI, my question is: If I were to increase the chunking size to something larger, such as 2000, would that present any issues? Would it increase or decrease the performance of the model? I'm curious to understand the tradeoffs and implications of using a larger chunking size. Does it have an impact on things like model accuracy, training time, or memory usage? I'd appreciate your insights on how the chunking size can affect the overall performance and effectiveness of the machine learning system. Thank you again for your valuable contributions. I look forward to hearing your thoughts on this follow-up question.

@markshark337 2 ай бұрын

Awesome, appreciate you're sharing it 👍does Langflow support other inputs from text, for example images, tables, etc. does it have OCR komponent supported too ?

@rsivakanth 2 ай бұрын

Great video @svpino, as always. However, I found that the python code using sample.json doesn't provide accurate results always as we get from the 'Run' using the browser. Even the same question that works well on browser gets a totally unrelated result on Python code when run with the json. Do you suggest anything on this?

@alextiger548 2 ай бұрын

Excellent man! How i can use SQL database for the RAG instead of the URL to provide the context?

@durgabhavani7860 9 күн бұрын

I can not find vector search component in lang flow. Could you please tell us Which version are you using? Or alternate solution to it

@fuzzylop1857 5 күн бұрын

I am using LangFlow v1.0.6 and there is no "Vector Search" 😞

@artur50 2 ай бұрын

how can you set it up with Ollama?

@AdandKidda Ай бұрын

in langflow , can we have "RAG + Tool usage" , at the same time.?

@alextiger548 2 ай бұрын

how to use the keys from .env to run the python code if i import your flow json without embedded keys

@underfitted 2 ай бұрын

Use the python-dotenv library.

@mikeg3810 2 ай бұрын

I need an app where I can transcribe and summarize videos, audio files and from sites like KZfaq similar to an app like Eightify or Summarify.

@solomonpaul6279 2 ай бұрын

Does this took work offline? Meaning it doesn't effect our data privacy please?

@ezioweb 14 күн бұрын

Can I use newer Python version (like 3.11.3)?

@underfitted 13 күн бұрын

I think yes, but you should try

@coderider3022 28 күн бұрын

My input components missing , why is that ?

@africoursity 2 ай бұрын

It seems to me that your interface looks different from mine. I don't have a core component but several other menus which I don't see in yours. What have I done wrong? Great tutorial, Man.

@underfitted 2 ай бұрын

Make sure you run the prerelease version

@guygrundman7969 2 ай бұрын

the URL i am working with requires login, how would it be handled in this case?

@underfitted 2 ай бұрын

Can’t do it like this. You need to find a way to authenticate and get access to the information.

@nigereast3796 Ай бұрын

Try agents .. crewai

@morphos2 2 ай бұрын

👌👏👏👏👏👏👏👏👏

@Omnicypher001 2 ай бұрын

ok, now how do you verify the security of the app? how do you make sure it isn't stealing your customers data?

@nigereast3796 Ай бұрын

Use it for POC only .. for production that different

@NoCommentLeftUnsaid 2 ай бұрын

something has changed since the video, it's taken over 6 hours and the install is still running Downloading wcwidth-0.2.9-py2.py3-none-any.whl (102 kB) |████████████████████████████████| 102 kB 4.2 MB/s