Ollama with Vision - Enabling Multimodal RAG

Поділитися
Вставка
  • Опубліковано 28 лис 2024

КОМЕНТАРІ • 16

  • @throwaway-g9f
    @throwaway-g9f 21 день тому +4

    This video got me hyped; I was waiting for ollama multi-modal for a long time.

  • @TheRealHassan789
    @TheRealHassan789 21 день тому +3

    this video and tools have so much value! ...people will sleep on it tho

  • @samsquamsh78
    @samsquamsh78 21 день тому +1

    great video and cool project! will chekc it out!! thanks!

  • @HappyDancerInPink
    @HappyDancerInPink 21 день тому +1

    Nice, what GPU do you use for these tests?

    • @engineerprompt
      @engineerprompt  21 день тому +2

      I have a MacBook Pro M2 Max with 96GB unified memory

  • @Masoud2xm
    @Masoud2xm 14 днів тому

    I am getting an error during indexing saying "Torch not compiled with CUDA enabled". I am using Mac M4. Could you help with this, please?

    • @timstevens3361
      @timstevens3361 12 днів тому

      get an rtx 3060 12 gig gpu or rtx 4060 16 gig
      they run alot of diff model really well !!!

  • @ChristopherMcKinley-c1s
    @ChristopherMcKinley-c1s 21 день тому

    Is there api calling in the future for this project? I would love to be able to use it as a replacement/upgrade from fine tuning models and running them from ollama.

    • @truthwillout1980
      @truthwillout1980 21 день тому

      ???

    • @ChristopherMcKinley-c1s
      @ChristopherMcKinley-c1s 21 день тому

      @@truthwillout1980 The idea in my head is that I can host this on the LAN and have other programs just make an api call so as to not go through a GUI. Is that already an option and I missed it?

    • @truthwillout1980
      @truthwillout1980 21 день тому

      @@ChristopherMcKinley-c1s Yes you should already be able to do that. I think there's a section in the video that explains it in fact (though I'm going off memory, I haven't watched it again). Just spin it up on a port number and call it.

  • @faucetcryptos8148
    @faucetcryptos8148 21 день тому +1

    Muito Legal

  • @Know_Ur_World
    @Know_Ur_World 20 днів тому

    So can u help me with the usecase
    My usecase it to extract the relevant text and images available in the pdf.when any prompt is given then relevant text alogh with image should display as reponse in a sequential manner, not images separate text separate.
    Query:Give steps in RSA agent installation
    Answer:
    1.Text1
    Image1
    2.Text2
    3.Image2
    Text 3
    4.Image4
    Text4
    5.image5
    Image 6
    Text5

  • @stunspot
    @stunspot 21 день тому

    Neat!

  • @mr.gk5
    @mr.gk5 21 день тому

    Can it generate graphs or reports on tabular data?

    • @engineerprompt
      @engineerprompt  21 день тому +1

      At the moment, it can't but I think its possible to integrate it with a code interpreter for plots or table generation.