A hyper-fast local vector database for use with LLM Agents • Hacker News Board

Re: A hyper-fast local vector database for use with LLM Agents

#21

Post by 0x008 » Mon, Apr 17, 2023, 6:28 AM UTC

kristianp wrote: post #20

egeozcan wrote: post #7

https://github.com/jdagdelen/hyperDB/blob/main/hyperdb/galax... This was a great laugh. Also, after the benchmark, it says: > Benchmark Credit: Benim Kıçım "Benim Kıçım" means "my ass" in Turkish.

References "[Submitted on 31 Mar 2023] Programmed but Arbitrary Control Minimization of Amplitude and phase for speckle Nulling (PACMAN)" https://arxiv.org/abs/2303.18231 No idea if that's a joke paper.

> HyperSVMRanking (Such Vector, Much Ranking) algorithm proposed by Andrej Karpathy (2023) https://arxiv.org/abs/2303.18231

Re: A hyper-fast local vector database for use with LLM Agents

#22

Post by elforce002 » Mon, Apr 17, 2023, 11:14 AM UTC

celestialcheese wrote: post #11

Damn this is good - well played. For those who aren't neck deep in LLM hype - ChromaDB raised $18m at a $75m valuation a couple weeks ago with what's essentially just a thin wrapper around duckdb with a parquet import/export for persistence.

Hahaha. Really?! Good for them.

Re: A hyper-fast local vector database for use with LLM Agents

#23

Post by jdaggers » Mon, Apr 17, 2023, 3:33 PM UTC

yawnxyz wrote: post #15

aside from making fun of Chroma, this seems to 100% work though? What would be the pros/cons of actually using this instead?

Author here. Yes, it actually does work. If I’m ever indexing fewer than ~100k documents/paragraphs I would use something like this approach rather than reaching for a vector database.

Re: A hyper-fast local vector database for use with LLM Agents

#24

Post by gavi » Tue, Apr 18, 2023, 6:27 PM UTC

alfor wrote: post #13

It’s a joke, but is it a functional one? I found it crazy to run auto-GPT with an external DB over the internet, how can this be ever fast and efficient? (unless the dataset is humongous)

Yes, it used OpenAI embeddings and numpy to find cosine similarity

Re: A hyper-fast local vector database for use with LLM Agents

#25

Post by gavi » Tue, Apr 18, 2023, 6:27 PM UTC

jdaggers wrote: post #23

yawnxyz wrote: post #15

aside from making fun of Chroma, this seems to 100% work though? What would be the pros/cons of actually using this instead?

Author here. Yes, it actually does work. If I’m ever indexing fewer than ~100k documents/paragraphs I would use something like this approach rather than reaching for a vector database.

Great work!