The Best RAG Technique Yet? Anthropic’s Contextual Retrieval Explained!

Поделиться
HTML-код
  • Опубликовано: 10 ноя 2024

Комментарии • 104

  • @tvwithtiffani
    @tvwithtiffani Месяц назад +132

    For anyone wondering, I did try these methods (contextual retrieval + reranking) with a local model on my laptop. It does work great the rag part but it takes a while to import new documents due to chunking, generating summaries and generating embeddings. Re-ranking on a local model is surprisingly fast and really good with the right model. If you're building an application using rag, I'd suggest you make adding docs the very first step in the on-boarding to your application because you can then do all of the chunking etc in the background. The user might be expecting real-time drag->drop->ask question workflow but it wont work like that unless you're using models in the cloud. Also, remember to chunk, summarize and gen embeddings simultaneously, not one chunk after another as of course that'll take longer for your end-user.

    • @kenchang3456
      @kenchang3456 Месяц назад +2

      Thanks for the follow-up.

    • @TheShreyas10
      @TheShreyas10 Месяц назад +6

      Can you share the code if possible

    • @ashwinkumar5223
      @ashwinkumar5223 Месяц назад +2

      Nice

    • @ashwinkumar5223
      @ashwinkumar5223 Месяц назад +1

      Will you guide to do the same

    • @tvwithtiffani
      @tvwithtiffani Месяц назад

      @@ashwinkumar5223 Unfortunately I cannot share code but I can advise. Just remember that everything runs locally. The language model, the embeddings model (very small compared to llm), the vector db (grows in GBs as you add more docs. This is where the generated embeddings are labeled & stored). A regular db for regular db crud stuff & keeping track of the status of document processing jobs. I went with mongodb because its a simple nosql data store that has libraries and docs for many programming languages. These dbs and models are ideally held in memory, but for resource constrained systems, you may want to orchestrate the loading and unloading of models as needed during your workflow. How would depend on the target platform you're developing for, desktop vs native mobile, vs web. I say all of this to say make sure you have a lot of system RAM and hard drive space. Mongo recently added some support for vectors given the noise around llms lately so there may be a bit of overlap here. But I haven't checked it out. Might not need a vectordb AND mongodb...

  • @BinWang-b7f
    @BinWang-b7f Месяц назад +162

    Sending my best to the little one in the background!

  • @seanwood
    @seanwood Месяц назад +2

    Working with this now and didn’t use the new caching method 😫. Nice to have someone else run through this 🎉😆

  • @megamehdi89
    @megamehdi89 Месяц назад +36

    Best wishes for the kid in the background

  • @tomwawer5714
    @tomwawer5714 Месяц назад +1

    Thanks very interesting. Many ideas came to my head for improving RAG with enhancing chunk

  • @SunilM-x9o
    @SunilM-x9o 11 дней назад +1

    what if the document is so big, that it couldn't fit in the llm context window how do we get the contextual based chunks then.
    if we consider break the document into small segments/documents to implement this approach, won't it lose some context with it

  • @IAMCFarm
    @IAMCFarm Месяц назад +4

    Applying this to local models for large document repos seems like a good combo to increase RAG performance. I wonder how you would optimize for the local environment.

  • @vikramn2190
    @vikramn2190 Месяц назад +2

    Thanks for the easy to understand explanation

  • @MatichekYoutube
    @MatichekYoutube Месяц назад +7

    do you maybe know what is going on in GPT Assistants - cause they rag is really efficiant - accurate - they have default 800 token chunks and 400 overlap. And it seems to work really well.Perhaps they use somekind of re-ranker also? Maybe you know ..

  • @alexisdamnit9012
    @alexisdamnit9012 Месяц назад

    Great explanation 🎉

  • @stonedizzleful
    @stonedizzleful Месяц назад

    Great video man. Thank you!

  • @yt-sh
    @yt-sh Месяц назад

    really useful article and video!

  • @i2c_jason
    @i2c_jason Месяц назад +8

    Hasn't structured graphRAG already solved this? Find the structured data using a graph, then navigate it to pull the exact example?

    • @remusomega
      @remusomega Месяц назад

      How do you think the Graph gets structured in the first place

    • @faiqkhan7545
      @faiqkhan7545 Месяц назад +2

      @@remusomega Any Links to read ?

    • @MyBinaryLife
      @MyBinaryLife Месяц назад

      @@faiqkhan7545 checkout the microsoft graphrag repo, pretty useful

    • @MyBinaryLife
      @MyBinaryLife Месяц назад

      @@faiqkhan7545 check out the microsoft graphrag repository

  • @DRMEDAHMED
    @DRMEDAHMED Месяц назад +1

    I want to add this as a the default way the rag is handled in open webUI but its conflicting with other stuff, I tried to make a custom pipeline for it but i'm struggling to make it work is it out of the scope of open web UI or am I just not understanding the documentation properly

  • @jackbauer322
    @jackbauer322 Месяц назад +13

    I think the baby in the background disagrees :p

  • @PeterJung-cx1ib
    @PeterJung-cx1ib Месяц назад

    How is the diagram generated/built at 0:48 for RAG embeddings?

  • @limjuroy7078
    @limjuroy7078 Месяц назад +2

    What happened if the document contains a lot of images like tables, charts, and so on? Can we still chunk the document in a normal way like setting a chunk size?

    • @kai_s1985
      @kai_s1985 Месяц назад

      You can use vision based rag, he described in his previous video.

    • @limjuroy7078
      @limjuroy7078 Месяц назад

      @@kai_s1985, so we don't need to chunk our documents if we use vision based RAG? My problem is how are we going to chunk our documents even though the LLM has vision capabilities

    • @kairatsabyrov2031
      @kairatsabyrov2031 Месяц назад

      @@limjuroy7078 it is very different from the text based rag. But, I think you need to embed images page by page. Look at his video or read the ColPali paper.

    • @awakenwithoutcoffee
      @awakenwithoutcoffee Месяц назад

      @@limjuroy7078 no, you would still need to chunk/parse your PDF's into text/tables/extracted images -> store those in 2 separate databases (s3/blob storage for images ) -> embed the images and the text separately -> on Query retrieve the closest images/text from these 2 stores in parallel -> feed to the OCR Model which will analyze the context including texts & image(s).
      There are more ways to use Vision models though: ColPali is one of them that is discussed by the OP in a different video. The approach here is to directly embed each page of a PDF/source as a picture and embed them directly. It's an interesting approach but with several drawbacks that source content isn't extracted/stored/accessible directly for queries/analysis but only at run-time. To get insight in your data you would need an OCR model to process the pages directly.

  • @loudcloud1499
    @loudcloud1499 Месяц назад

    very informational visualizations!

  • @AlfredNutile
    @AlfredNutile Месяц назад

    Great work!

  • @souvickdas5564
    @souvickdas5564 Месяц назад +1

    How to generate those context for chunks without having the sufficient information to the LLM regarding the chunk? How they are getting the information about the revenue numbers in that example? If it is extracted from the whole document then it will be painful for llm cost.

    • @zachmccormick5116
      @zachmccormick5116 Месяц назад +1

      They put the entire document in the prompt for every single chunk. It’s very inefficient indeed.

    • @karthage3637
      @karthage3637 Месяц назад

      @@zachmccormick5116well it’s not inefficient if you can cache the prompt
      They find a way to push this feature

  • @samuelimanuel7643
    @samuelimanuel7643 22 дня назад

    I'm still new learning about RAG, but want to ask how would this differ or fit it with graphRAG? I heard GraphRAG are really well connected?

  • @andrew-does-marketing
    @andrew-does-marketing Месяц назад +1

    Do you do contract work? I’m looking to get something like this created.

    • @engineerprompt
      @engineerprompt  Месяц назад +1

      Yes, you can contact me. Email is in the video description.

  • @konstantinlozev2272
    @konstantinlozev2272 Месяц назад

    Losing the context in RAG is a real issue that can destyall usefulness.
    I have read that a combination of the chunks and Graphs is a way to overcome that.
    But have not tested with a use case yet myself.

    • @NLPprompter
      @NLPprompter Месяц назад

      I'm interested why graph can be useful for LLM to able retrieve better

    • @konstantinlozev2272
      @konstantinlozev2272 Месяц назад +1

      @@NLPprompter My understanding is that graphs condense and formalise the context of a piece of text.
      My use case is a database of case law.
      There are some obvious use cases for that when a paragraph cites another paragraph from another case.
      But beyond that I think there is a lot of opportunity is just representing each judgement in a standardised hierarchical format.
      But I am not 100% sure how to put all together from a software engineering perspective.
      And maybe one could use relational database instead of graphs too.🤔

    • @NLPprompter
      @NLPprompter Месяц назад +1

      @@konstantinlozev2272 graph indeed is fascinating, maybe I'm not really know what and how it's able related to LLMs, what's makes it interesting is when Grokking state happen and model reach to be able generalize it's training, they tend to create a pattern with their given data, and those pattern are mostly geometric patterns, really fascinating although i tried to understand that paper which i can't comprehend with my little brain.... so i do believe graph rag somehow also have meaning/useful for llm.

    • @konstantinlozev2272
      @konstantinlozev2272 Месяц назад +1

      @@NLPprompter I guess it will have to be the LLM working with the API of the knowledge graph which function calling

  • @CryptoMaN_Rahul
    @CryptoMaN_Rahul Месяц назад

    I'm working on AI POWERED PREVIOUS YEAR QUESTIONS ANALYSIS SYSTEM WHICH WILL ANALYZE TRENDS AND SUMMARY OF PREVIOUS 5-10 YEARS PAPER AND WILL GIVE A DETAILED REPORT OF IMPORTANT TOPICS ETC.. can you tell what should be the approach to implement this ?

  • @ashutoshdongare5370
    @ashutoshdongare5370 29 дней назад

    How this compare with Graph Hybrid RAG ?

  • @wwkk4964
    @wwkk4964 Месяц назад +4

    🎉baby voices were cute!

  • @RedCloudServices
    @RedCloudServices Месяц назад

    Do you think Visual LLMs like ColPali provide accurate context and results than traditional RAG using text-based LLMs?

  • @janalgos
    @janalgos Месяц назад +2

    how does it compare to hybridRAG?

  • @martinsherry
    @martinsherry Месяц назад

    V helpful explanation.

  • @LatifAmars
    @LatifAmars Месяц назад

    What tool did you use to record the video?

  • @steve-g3j6b
    @steve-g3j6b Месяц назад

    @Prompt Engineering didnt find a clear answer for my question, so I ask you. as a screenplay writer what do you think is the best model for me? gpt has very short memory. (not enough token memory)

    • @kees6
      @kees6 Месяц назад

      Gemini?

    • @steve-g3j6b
      @steve-g3j6b Месяц назад

      @@kees6 why?

    • @lollots82
      @lollots82 Месяц назад

      ​@@steve-g3j6bhas had 1M token window for a while

  • @SonGoku-pc7jl
    @SonGoku-pc7jl Месяц назад +1

    thanks!

  • @udaym4204
    @udaym4204 Месяц назад

    does Multi-Vector Retriever Worth It?

  • @VerdonTrigance
    @VerdonTrigance Месяц назад

    How did they put a whole doc into prompt?

    • @vaioslaschos
      @vaioslaschos Месяц назад

      most commercial LLm have a window of 120k or more. But even if this not the case, you can just take much bigger chunks as context.

  • @nealdalton4696
    @nealdalton4696 Месяц назад

    Are you adding this to localGPT?

  • @HawkFranklinResearch
    @HawkFranklinResearch Месяц назад

    Contextual retrieval just seems equivalent to GraphRag (by Microsoft) that indexes knowlegde context wise

  • @robrita
    @robrita Месяц назад +15

    can hear baby in the background 👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶👶

  • @loicbaconnier9150
    @loicbaconnier9150 Месяц назад +4

    I you want to make, the embedding, bm25 and reranker , just use Colbert it's more efficient...

    • @the42nd
      @the42nd Месяц назад +1

      True, but he does mention colbert at 09:45

    • @engineerprompt
      @engineerprompt  Месяц назад +2

      ColBERT is great but there are two major issues with it currently, which hopefully will be addressed soon by the community.
      1. Most of the current vectorstores lack support for it. I think qdrant has added the support. Vespa is another one but the mostly used ones still need to add that support.
      2. The size and storage needs is another big issue with colbert. Quantization can help but I haven't seen much work on it yet.

    • @loicbaconnier9150
      @loicbaconnier9150 Месяц назад +1

      It’s very quick indexing documents, i use it as another retreiver in llamaindex. I create several index with it to improve or check retrieved chunks
      But you are right, best option to keep index is Qdrant.

  • @DayLearningIT-hz5kj
    @DayLearningIT-hz5kj Месяц назад

    Love the Baby ❤️ good father !

  • @MrGnolem
    @MrGnolem Месяц назад

    Isn't this what llama index has been doing for over a year now?

  • @crashandersen602
    @crashandersen602 Месяц назад

    So easy a baby could do it. Don't believe us? We have one following along in this lesson!

  • @ibrahimaba8966
    @ibrahimaba8966 Месяц назад

    This is the best way to sell their features: prompt caching 😁.

  • @MrAhsan99
    @MrAhsan99 Месяц назад

    You can name the little one "Ahsan" just in case, if you are looking for the names.

  • @micbab-vg2mu
    @micbab-vg2mu Месяц назад

    interesting :)

  • @NLPprompter
    @NLPprompter Месяц назад

    so we are going to have chunking model, embedding model, graph model, and conversation model... and they can work within program called by lines of codes, or... they can work freely fuzzyly in agentic way...
    i imagine a UI of game dev, drag and drop pdf to them, they will busy working on that file running around like cute little employee, and when done user can click a pc item then it will.... ah nevermind that would be waste of VRAM

  • @isaacking4555
    @isaacking4555 Месяц назад

    The baby in the background 🤣

  • @marc-io
    @marc-io Месяц назад +2

    so nothing new really

  • @LukePuplett
    @LukePuplett Месяц назад

    I was so astonished by how obviously terrible the original "dumb chunking" approach is that I couldn't watch the video.

  • @finalfan321
    @finalfan321 Месяц назад +1

    you sound tired but i thin i know why ;)

  • @karansingh-fk4gh
    @karansingh-fk4gh Месяц назад

    Your voice is very low. So difficult to understand entire things

  • @jensg8547
    @jensg8547 Месяц назад

    Vector embedding solutions for retrieval are doomed as soon as SLMs get cheap and fast enough. Why relying on cosine similarity when you can instead query a llm over all search data at inference time?!

  • @cherepanovilya
    @cherepanovilya Месяц назад

    old news))

  • @yurijmikhassiak7342
    @yurijmikhassiak7342 Месяц назад +1

    WHY NOT TO DO SMARK CHANKING ON CONTENT. LIKE WHEN NEW TOPIC STARTS? NEW SENTENCE, ETC? YOU WILL USE FAST LLM TO GENERATE CHANKS. THERE WILL BE LESS NEED FOR OVERLAP.

    • @Autovetus
      @Autovetus Месяц назад

      Chill , dude... Sheesh 🙄

  • @hayho4614
    @hayho4614 Месяц назад

    maybe speaking with a bit more energy would keep me more engaged

  • @snapman218
    @snapman218 Месяц назад +1

    Good information, but having a child crying in the background is unprofessional. Of course now everyone will say I hate children, but I don’t care. I’m sick of unprofessional behavior.

    • @kerbberbs
      @kerbberbs Месяц назад +2

      Its youtube dawg, nobody cares. Just watch the overlengthed vid and move on. Most people here only came for 2 mins of what's actually important

    • @ogoldberg
      @ogoldberg Месяц назад +3

      Rude thing to say, and ridiculous. You are the one who is unprofessional.

    • @tombelfort1618
      @tombelfort1618 Месяц назад +2

      Entitled much? How much did you pay him for his time again?

  • @ZukunftTrieben
    @ZukunftTrieben Месяц назад +2

    00:30