The Best, Ultra-Fast and Low-Cost GraphRAG: LightRAG

Поделиться
HTML-код
  • Опубликовано: 22 янв 2025

Комментарии • 26

  • @SridharKumarKannam
    @SridharKumarKannam  3 месяца назад +3

    If you found this content helpful, please consider liking, subscribing, and sharing it with others who might benefit. Your support is greatly appreciated :)

  • @ronifintech9434
    @ronifintech9434 Месяц назад

    Thank you for the knowledge sharing

  • @tonyseno
    @tonyseno 2 месяца назад +3

    Can the LightRAG show the references including the page numbers it refers to?

    • @SridharKumarKannam
      @SridharKumarKannam  2 месяца назад

      It doesn't have that feature out of the box, but the source code can be modified to save page number info as metadata for the nodes and can be returned to the user. Its not an easy task though...

  • @sndrstpnv8419
    @sndrstpnv8419 9 дней назад

    do yo have your code project how to use LightRAG

  • @siddharthvj1
    @siddharthvj1 2 месяца назад

    love this channel

    • @SridharKumarKannam
      @SridharKumarKannam  Месяц назад

      Thank you very much. Please consider sharing it with others who might benefit. Your support is greatly appreciated :)

  • @kartikv776
    @kartikv776 3 месяца назад

    Love these updates!

    • @SridharKumarKannam
      @SridharKumarKannam  3 месяца назад

      If you found this content helpful, please consider liking, subscribing, and sharing it with others who might benefit. Your support is greatly appreciated :)

  • @stonkmaster6969
    @stonkmaster6969 2 месяца назад +1

    Is LightRAG's codebase not a little unstable for enterprise use? Also, it seems to have limitations with LLM model selection; supports only OpenAI models and Claude.

  • @SandeepS-i4e
    @SandeepS-i4e 2 месяца назад

    Can it be used for pdfs or douments containing images and tables also?
    Could you pls do a video on that also?

  • @薛帅-p7y
    @薛帅-p7y 3 месяца назад

    it can be used in producation envs statbly?

    • @SridharKumarKannam
      @SridharKumarKannam  3 месяца назад

      yes, the source code is available, just make sure there are no security issues related to API call to the servers...

  • @pill
    @pill 3 месяца назад

    I wonder how this would have to be changed in order to work over a codebase. I have a use case of natural language to Python and am trying traditional RAG, but I think a graph would help. Do you know of any other options?

    • @SridharKumarKannam
      @SridharKumarKannam  3 месяца назад

      The codebase is tricky. A couple of suggestions.
      1. Write docstrings for all the functions and important and complex code lines.
      2. Use LLMs to create functions decryption.
      3. Use the description and docstrings info to create embeddings and add them to the graph entity nodes.
      4. Use such info in the retrieval.
      All the best :)

  • @icanride4long
    @icanride4long 3 месяца назад

    Hi, great content. Easy to follow. Thanks for that. Are you able to compare and contrast LightRag vs GraphReader? GraphReader is similar in flow to GraphRag so I assume it has the same disadvantages. But would love to hear your thoughts since you've posted content on both solutions. Which would you recommend for production, today?

    • @SridharKumarKannam
      @SridharKumarKannam  3 месяца назад

      thanks, I'll try but i'm occupied with quite a few things at the moment.

  • @AbdulAziz-my3wt
    @AbdulAziz-my3wt 3 месяца назад

    thanks, it is great, can we change chunk size?

    • @SridharKumarKannam
      @SridharKumarKannam  3 месяца назад

      yes, you can when instantiating the RAG
      rag = LightRAG(
      working_dir=WORKING_DIR,
      chunk_token_size = N
      )

    • @AbdulAziz-my3wt
      @AbdulAziz-my3wt 2 месяца назад

      Thanks can i do hashing to prevent reindexing on same documents?

  • @piyushaaryan869
    @piyushaaryan869 3 месяца назад

    Is this rag can be used for production ?

    • @SridharKumarKannam
      @SridharKumarKannam  3 месяца назад

      yes, the source code is available, just make sure there are no security issues related to API call to the servers...

  • @georgetorres1535
    @georgetorres1535 3 месяца назад

    Is it possible to change your vector database? Or will there be problems?

    • @SridharKumarKannam
      @SridharKumarKannam  3 месяца назад

      You can change the vector DB from nano-vectordb in the source code.