USING 🦜 LANGCHAIN AIs SequentialChain 🔗 : BUILD -

[@Avra_b]

You can find the code and more in the blog post I published - Link - medium.com/@avra42/getting-started-with-langchain-a-powerful-tool-for-working-with-large-language-models-286419ba0842 Let me know what do you all think . If you like the video, don't forget to like , share and subscribe , Cheers 🤗

[@ZapCrafter]

Thanks

[@Avra_b]

Hey Paul ! Very kind of you . Thanks 🤗

[@bg-mq5hz]

Thanks!

[@Avra_b]

Such a kind gesture , thank you for your contribution 🤗

[@1littlecoder]

Great stuff! Thanks for covering the latest hot topic!

[@Avra_b]

Thanks 😊 Coming from you , means a lot 🤗💙

[@daniiielsan]

Long time looking for a good Langchain explanation and this is the best one by far. Thanks Avra

[@Avra_b]

Hi Daniel ! Very happy to hear such positive feedback. I’m glad you liked the video. Stay in touch 🤗

[@VedicVisionsToday]

Good stuff , thanks for keeping it short and simple !

[@Avra_b]

Glad to hear 🤗

[@DEEPAKSV99]

Thank you so much for explaining this relatively new topic, in such simple terms. Your channel has a lot of potential to thrive in the near future.

[@Avra_b]

That’s very kind of you Deepak :) happy to hear such positive feedbacks. Stay in touch 🤗

[@andfanilo]

😍 loved it, very comprehensive!

[@Avra_b]

Thank you Fanilo! 😊

[@JayAditya-ih5zy]

Thankyou 😁

[@Avra_b]

You’re welcome 😊

[@ObservingBeauty]

Thanks! Are you familiar with LangFlow ? (Langchain implemented over react-flow)

[@Avra_b]

Heard about it over Twitter (if I'm not wrong) . But haven't tried it yet . Did you ? Would be nice if you share your experience - I'm curious.

[@ObservingBeauty]

@@Avra_b I wish I knew how :-)... not a coder, hence the search for more graphic interfaces. Thanks

[@Avra_b]

@@ObservingBeauty I would check it out . If it's a common interest would certainly make a tutorial video. Thanks for the suggestion :)

[@ObservingBeauty]

@@Avra_b fantastic

[@PabloRosa-y7p]

is it possible to break a SequentialChain? Here's my problem: The first chain must return True or False, if True run al chain. If False, break at the first link. Of course, using Langchain. Thanks guys!

[@Avra_b]

I think you can tweak the prompt to do so ( haven’t tried it though ) . Let me know if it works !

[@joesfilms1334]

Hey! I'm having a problem when programming a web app in python, I'm using streamlit, associated with firebase to perform the login process. After logging in, some buttons appear on the screen that perform certain functions, but when pressing the button, instead of performing its proper function, everything the user receives after login disappears, as if the user was not logged in. Can anybody help me? I can send the lines of code and a video of the problem if needed, thanks!

[@Avra_b]

Thanks for checking with your issue. From what I understand, this is mainly because of the Streamlits default workflow ( top to bottom ) . You need to keep the buttons state in active - you can use session state for that or just replace button with a checkbox . ( detail tutorial video here - kzbin.info/www/bejne/moHHc2qwrqynqsk )

[@AlgorithmicEchoes]

Can you make a video on how to recreate the same sequential chaining using the new GPT-3.5-Turbo endpoint in langchain? I have not found any tutorial on stramlit+ChatgptTurbo being used for text completion tasks (especially sequential chaining)

[@Avra_b]

Would you mind checking my latest video ( also a blog post is there ) . I’ve used GPT3.5 for text completion ( despite the fact I agree it’s ideal use case can be chat ) . You can just tweak the code Accordingly and use for this same use-case . kzbin.info/www/bejne/mXnNnaObp6qGmMk Code / blog pinned in the comment section . Let me know if it works :)

[@AlgorithmicEchoes]

@@Avra_b Great. Thanks! I'll check it out. I feel that the completion api is very good to build apps. Hopefully they'll give a gpt-4 completion endpoint and not just a chat endpoint.

[@Avra_b]

@@AlgorithmicEchoes I'm not sure here, but the latest GPT-3.5-Turbo, serves both the purpose right ? Completion and chat ? That's what I found in their official doc. Do you think, this would make a lot of differences ?

[@AlgorithmicEchoes]

@@Avra_bThank you for your enthusiastic response. I'm new to this so it could be that I'm mistaken but I'd like to explain my confusion. I think it stems from the way langchain handles llm and chat endpoints differently. I made an app based on your video of sequential chaining where the output of first chain is input of second and so on. Now here, in a bot like interface the prompt structure and prompt templates are different for langchain (and for chain 2, I'm unsure exactly what prompt should I pass?) To be specific, in the prior case you gave a statement and chain 1 verified its authenticity while chain 2 listed the assumptions. Here, I'm unsure this will happen automatically in the case of a conversationalist bot. I'll need to give a text input for chain 2, prompting it to question the assumptions). I'll try to figure it out and If I do, ill surely write to you.

[@Avra_b]

@@AlgorithmicEchoes Thanks for your inputs. I must say, you have quite a in-depth knowledge . Tbh, this is a very new implementation (LangChain + LLM models ) - still a black box at least for me in specific use-cases . The best way is to test & trial, as you suggested . However, this particular tweet might interest you - twitter.com/JAPozueloM/status/1636944289337942018?s=20 Dependency of model matters (you are right previously - if that's what you meant). Please do carry out experimenting, and do share with me and the community - it would be of huge help for everyone ! Cheers and Goodluck :)

[@FelipeRocha-y2u]

Hey sir, is it possible to use langchain + framework or library js? Like React or React Native

[@Avra_b]

For sure , LangChain has JS library . Check it out in their official repo or a quick search . Good luck 🍀

[@klammer75]

Well done sir! I love this and thank you for the information so eloquently presented🤓🥳🦾

[@Avra_b]

Glad you enjoyed it!