Building agentic LLM application Workflows with LlamaIndex

Поділитися
Вставка
  • Опубліковано 11 гру 2024

КОМЕНТАРІ • 17

  • @Last-Code-Bender
    @Last-Code-Bender 4 місяці тому +2

    What are some use cases for having multiple @steps functions that accept the same events. Additionally, do these @step functions that accept the same event run in parallel when the event is broadcast?

  • @VLM234
    @VLM234 4 місяці тому +2

    @LlamaIndex, It looks like similar to langgraph? Why one should choose workflows over langgraph??

  • @IdPreferNot1
    @IdPreferNot1 4 місяці тому

    Looks awesome. Can we get a little auto-help with these 27 different imports to remember each time a new llama-index feature comes out?

  • @kishorkukreja7733
    @kishorkukreja7733 3 місяці тому

    One suggestion, can we have a functionality like Collect Events that would check not just the type of events but also ensure that they are executed in the right order ?

  • @reynoldoramas3138
    @reynoldoramas3138 4 місяці тому +3

    I don´t get the point of this. Why I would use this instead of simple and normal conditions and loops? What is the aggregated value?

    • @xuexileader
      @xuexileader 3 місяці тому

      I guess cause the async better.

  • @ahsanmasood1459
    @ahsanmasood1459 4 місяці тому +1

    What exactly events are referring to? startevent, endevent etc

    • @lloganm1234
      @lloganm1234 4 місяці тому

      Events are just events. The trigger some step to execute, and can also hold extra information to execute that step.
      The user defines all events themselves, except the StartEvent (which signifies the entry point) and the StopEvent (which stops the workflow)

  • @Pingu_astrocat21
    @Pingu_astrocat21 4 місяці тому

    how to use hugging face models with this?

  • @pavankumarmantha
    @pavankumarmantha 4 місяці тому

    what is the difference between the utils "draw_all_possible_flows" and core "draw_all_possible_flows" ?

    • @LlamaIndex
      @LlamaIndex  4 місяці тому

      It was originally in core and has been moved to utils.

  • @sarthakkaushik17
    @sarthakkaushik17 4 місяці тому +2

    Is'nt it same as Langraph ?

    • @riteshsingh811
      @riteshsingh811 2 місяці тому +1

      Personally I really like langgraph implementation. It’s much more readable than this .

  • @jenniferdsouza7708
    @jenniferdsouza7708 2 місяці тому +1

    I am curious why OpenAI GPT models were used? On a different and somewhat related note, does LlamaIndex support calling various LLMs including Llama and other open-source kinds?

  • @KumR
    @KumR 4 місяці тому +1

    I think a better realistic example will make more sense

    • @lloganm1234
      @lloganm1234 4 місяці тому

      Lots of examples instead docs! This was meant to explain core concepts and highlight general design patterns workflows

  • @hbhavsi
    @hbhavsi 2 місяці тому

    It's 2024, we are blessed with AI, and people are putting out good videos on complex topics. But people still refuse to use a decent mic or watch what they themselves create.