Implementing RAG with SQL Database Vector Search and Semantic Kernel | Data Exposed: MVP Edition

Поділитися
Вставка
  • Опубліковано 16 гру 2024

КОМЕНТАРІ • 16

  • @CreativeEngineering_
    @CreativeEngineering_ Місяць тому

    I made a rag from scratch in c#. Its amazing!!! Vector database with lsh, real time memory index, partial match searching. All local using SQLite and fast Bert tokenizer.

  • @muralim8164
    @muralim8164 2 місяці тому +1

    This is Awesome!! Now I dont need Azure AI Search to upload my documents to do Vector Search, this reduies the plumping work of index creation, etc! Happy to see SQL database with vector support that could improve my search performance! 😎

  • @dsheardown
    @dsheardown 2 місяці тому

    Wow.. I have been playing with Azure AI for this type of thing, but Azure AI Search "can" be expensive... maybe not to a company, but as an individual. This is a really great way to achieve something similar without the extra overhead! thanks so much for sharing!

  • @JasonEvans01
    @JasonEvans01 2 місяці тому +2

    Nice useful solution. Thanks for sharing!

  • @rholdorf
    @rholdorf 2 місяці тому

    Exactly what I was looking for! Thanks!

  • @FotisMylonas
    @FotisMylonas 2 місяці тому

    Really helpful and inspiring, thanks!

  • @ClementFauchere
    @ClementFauchere Місяць тому

    What are the search performance? Do sort by distance does it scan the entire table or is there a way to use a clever index?

  • @bangonkali
    @bangonkali Місяць тому

    Hi. Can semantic kernel support other open source self hosed vector databases alternatives?

  • @gandharjayantaundhe4023
    @gandharjayantaundhe4023 2 місяці тому

    Great stuff! Thanks!

  • @LucianoEvaristoGuerche
    @LucianoEvaristoGuerche 2 місяці тому

    Hope more related videos to come since it caught my attention. Want to know whether it is possible to create index on the binary vector column (akin to geography indexes) or not; whether it is possible to use it with Semantic Kernel + EF to act as a vector store for long term memory and semantic indexing or not and if so, how.

  • @zfold4702
    @zfold4702 2 місяці тому

    Once upon a time Sql Server also was a Graph database. What happened to that? Who is using it in production or mission critical use cases?

  • @TheStoibz
    @TheStoibz 2 місяці тому

    How much of this is in production? I assume all preview at this stage and difficult to get on a tenant?

  • @LucianoEvaristoGuerche
    @LucianoEvaristoGuerche 2 місяці тому

    When you mentioned "SQL Server" on the title, do you mean Azure SQL, SQL server or both?

    • @heinrichmarais3753
      @heinrichmarais3753 2 місяці тому

      I'd also like to know!

    • @marcominerva
      @marcominerva 2 місяці тому

      At this moment, Vector Support is available in Early Adopter Preview on Azure SQL Database and Managed Instance.