Hey everyone, I hope you like the tutorial! Please note that I am using a small dataset in this example (just to get everything up and running) but you can scale to billions of vectors with pgvectorscale! You can check out the speed comparisons here: www.timescale.com/blog/pgvector-is-now-as-fast-as-pinecone-at-75-less-cost/
This couldnt have come at a better time for me. This is exactly what I needed right now. Such a great video/tutorial. Thank you so much for this and for sharing your code.
Great content! One benefit of some dedicated solutions is that integrating BM25 to utilize both sparse and dense vector fusion searches is built in - do you have a prefered method to accomplish this in postgres?
Great content! Thank you so much. I'm curious - why did you decide to implement the basic components on your own and not use one the more "familiar" frameworks - LangChain, LlamaIndex or Haystack?
Thanks! I don’t like relying on other framework when possible. They are usually too complex in my opinion. This way, there are limited external dependencies within the core functionality and I make sure I fully understand everything within the code base.
Hey everyone, I hope you like the tutorial! Please note that I am using a small dataset in this example (just to get everything up and running) but you can scale to billions of vectors with pgvectorscale! You can check out the speed comparisons here: www.timescale.com/blog/pgvector-is-now-as-fast-as-pinecone-at-75-less-cost/
Congrats on 100k. Another great video, creating quality rag systems is what I was learning about and this is very clarifying
Congrats on 100k subs, awesome tutorial
This couldnt have come at a better time for me. This is exactly what I needed right now. Such a great video/tutorial. Thank you so much for this and for sharing your code.
Awesome!
Thanks! For me, this is the best practical explanation of RAG. I will definitely use this approach in my projects.
Great content! One benefit of some dedicated solutions is that integrating BM25 to utilize both sparse and dense vector fusion searches is built in - do you have a prefered method to accomplish this in postgres?
Thanks I will try it:)
Great video Dave! Can you explain how to implement hybrid search on top of this i.e. integrate this together with a keyword-based search?
+1
I'd also like to see a video about BM25, Apache Calcite or elasticsearch - good keyword search is crucial for most use cases
Awesome content, thanks!
Glad you liked it!
Great content! Thank you so much.
I'm curious - why did you decide to implement the basic components on your own and not use one the more "familiar" frameworks - LangChain, LlamaIndex or Haystack?
Thanks! I don’t like relying on other framework when possible. They are usually too complex in my opinion. This way, there are limited external dependencies within the core functionality and I make sure I fully understand everything within the code base.