LangChain vs LangGraph: A Tale of Two Frameworks

Поділитися
Вставка
  • Опубліковано 21 лис 2024

КОМЕНТАРІ • 32

  • @abiofficial-ws7pn
    @abiofficial-ws7pn 18 днів тому +9

    I've said this before, and I'll say it again. His voice is soothing, and it's the reason I often watch his videos, even when I don't understand much of it

  • @ErroneousTheory
    @ErroneousTheory 18 днів тому +13

    What a way to start a Monday. I’ve been meaning to look into this

  • @amotorcyclerider3230
    @amotorcyclerider3230 18 днів тому +2

    Perfect. I was just getting started to look into langgraph and was looking for a video and this one was just perfect.

  • @tahirk9
    @tahirk9 4 дні тому

    Exactly what I was lookig for, Thanks!!

  • @throughiris8676
    @throughiris8676 18 днів тому +5

    The one thing I can think about is, "Is he writing in reverse?" If he's writing on the backside of this glass wouldn't he have to write backwards, like in a mirror, for it top appear normal to the camera? It can't be a post effect where you flip it for the video because then this hand and marker wouldn't match the writing. The writing looks too natural and fluid to be in reverse...or is it? I can't pay attention to anything else till I figure this out

    • @IBMTechnology
      @IBMTechnology  18 днів тому +4

      If you want to see how it's done, check out this behind the scenes video ua-cam.com/video/Uoz_osFtw68/v-deo.htmlfeature=shared

  • @kishanmod5857
    @kishanmod5857 7 днів тому

    Insightful as always 😊

  • @AleksandrSedov-o6o
    @AleksandrSedov-o6o 17 днів тому

    Nice format, great insights! Not many ways to get that..

  • @raptordudu
    @raptordudu 17 днів тому

    Can Laingraph do sequential tasks? If so why do we still need lainchain?

  • @tee_iam78
    @tee_iam78 6 днів тому

    nice video. thank you.

  • @GenAISpotlight
    @GenAISpotlight 17 днів тому

    Excellent video, thank you 👏

  • @anaghashinde1200
    @anaghashinde1200 12 днів тому

    How do u make these videos???

  • @NLPprompter
    @NLPprompter 12 днів тому

    wow cool video, kind to go langchain some more?

  • @leonardosilveira1297
    @leonardosilveira1297 18 днів тому

    Amazing as always!

  • @Minotaurus007
    @Minotaurus007 14 днів тому +1

    Hmm... While LangChain is easy to understand (great video btw), LangGraph is presented a little bit too abstract here. Perhaps some examples would have helped.

  • @engenglish610
    @engenglish610 15 днів тому +1

    It's free to use langraph?

    • @konrad-carrot-manager
      @konrad-carrot-manager 2 дні тому

      I too want to know! Is this a glass between camera and presenters? That would make the text and drawings reversed (mirrored). Or is this a mirror and the speakers are looking at the mirror? Then where is the camera in the reflection? :)

  • @SuperCatbert
    @SuperCatbert 16 днів тому

    very good

  • @teckyap7015
    @teckyap7015 6 днів тому

    Anyone here wants to know about real world applications which are using these mentioned framework? 😅

  • @markthien
    @markthien 18 днів тому +2

    how did this guy fraw on a transparent board? is this a web platform? anyone know? thanks bro!

    • @mao73a
      @mao73a 18 днів тому +2

      I suspect it's just a glass in front of the camera. In post processing they flip screen sides to make it readable for the audience.

  • @kevincitron2335
    @kevincitron2335 17 днів тому

    Well made vid.
    But. Show me some actual implementation examples.

  • @fuad471
    @fuad471 16 днів тому

    still not clear the difference of two frameworks in a practical manner. would need real-case examples to get the topic. For now cant decide which one to use when in what situation, can both tools handle same projects for ex ?

    • @АндрейПонамарев-ж2г
      @АндрейПонамарев-ж2г 16 днів тому

      because you'll still need to use agent (agent classes) from langchain while building the graph

    • @АндрейПонамарев-ж2г
      @АндрейПонамарев-ж2г 16 днів тому

      The best way to do it is to try it out. Build simple RAG locally or in Azure Portal with PostgreSQL vector DB and some simple agent (you can find ready to go examples in documentation) + add tools (description for agent on how and when to use this tool - ex. cover specific dataset; and prompt - to streamline further generation). It will cost you nothing, 2-3 days to play with it
      Real-case - me and my team build AI Assistant for an investment company. We store info about their investment products and help their clients to search for them or get recommendations. User asks the question regarding specific asset domain/sector. Agent analyzes the question, decides which data collection to retrieve from, which tool to use to process this data and gives everything to LLM which generates the final answer. This is langchain

    • @АндрейПонамарев-ж2г
      @АндрейПонамарев-ж2г 16 днів тому

      Talking about LangGraph - we can not only add new tools and prompts to cover different questions, but create workflows with several: tools /agent decisions /places to retrieve data/ switching between SQL or vector search.
      Besides that, there are various ways to augment or finetune the workflow (graph).

  • @engenglish610
    @engenglish610 15 днів тому

    N8N seems to be a joke compared to this.

  • @codewithbrogs3809
    @codewithbrogs3809 4 дні тому

    Uhhhh... that's not a correct explanation of chaining

  • @lordlee6473
    @lordlee6473 12 днів тому

    Not a really great explanation

  • @tahirk9
    @tahirk9 4 дні тому

    Exactly what I was lookig for, Thanks!!

  • @tahirk9
    @tahirk9 4 дні тому

    Exactly what I was lookig for, Thanks!!