Sharding Pgvector
(pgdog.dev)
If you find yourself working with embeddings, you’ve shopped around for a vector database. pgvector is a great option if you’re using Postgres already. Once you reach a certain scale (about a million arrays), building indices starts taking a long time. Some workarounds, like parallel workers, help, but you still need to fit the whole graph in memory.
If you find yourself working with embeddings, you’ve shopped around for a vector database. pgvector is a great option if you’re using Postgres already. Once you reach a certain scale (about a million arrays), building indices starts taking a long time. Some workarounds, like parallel workers, help, but you still need to fit the whole graph in memory.
Ask HN: Alternatives to Vector DB?
(ycombinator.com)
A while back I was looking for a vector database that would work across Windows / Mac / Linux platforms. Some of the options required specific processors like Intel. I am curious if there are any alternatives to a Vector DB that can run cross platform and are easy to setup?
A while back I was looking for a vector database that would work across Windows / Mac / Linux platforms. Some of the options required specific processors like Intel. I am curious if there are any alternatives to a Vector DB that can run cross platform and are easy to setup?
Why Vector DBs Are the Wrong Abstraction – and What We Built Instead
(topk.io)
We’ve spent the last three years building the most popular vector database on the market. In that time we realized that a database built around vectors as a primary key is simply the wrong abstraction, creating an unnecessary obstacle for users in production.
We’ve spent the last three years building the most popular vector database on the market. In that time we realized that a database built around vectors as a primary key is simply the wrong abstraction, creating an unnecessary obstacle for users in production.
Show HN: Tinyhnsw – The Littlest Vector Database
(github.com/jbarrow)
TinyHNSW is a tiny, simple vector database. It weighs in at a measly few hundred lines of code. It's built on a straightforward (but not fast) implementation of HNSW in Python with minimal dependencies. It has an associated set of tutorials that build up to understanding how HNSW works, and how you can build your own TinyHNSW.
TinyHNSW is a tiny, simple vector database. It weighs in at a measly few hundred lines of code. It's built on a straightforward (but not fast) implementation of HNSW in Python with minimal dependencies. It has an associated set of tutorials that build up to understanding how HNSW works, and how you can build your own TinyHNSW.
Powering AI RAG Applications with Vector Embeddings
(sambanova.ai)
As AI developers strive to build faster, more accurate and contextually relevant Retrieval Augmented Generation (RAG) systems, they face significant challenges in efficiently managing large-scale unstructured data and delivering fast, accurate responses.
As AI developers strive to build faster, more accurate and contextually relevant Retrieval Augmented Generation (RAG) systems, they face significant challenges in efficiently managing large-scale unstructured data and delivering fast, accurate responses.
Why HNSW is not the answer and disk-based alternatives might be more practical
(pgvecto.rs)
HNSW (Hierarchical Navigable Small World) has become the go-to algorithm for many vector databases. Its multi-layered graph structure and ability to efficiently navigate vector embeddings make it particularly appealing. However, despite its apparent advantages, HNSW may not be the optimal solution for large-scale and dynamic vector similarity search. In this blog post, we challenge the dominance of HNSW and explore why disk-based alternatives, such as IVF (Inverted File Index), might be more practical for massive datasets.
HNSW (Hierarchical Navigable Small World) has become the go-to algorithm for many vector databases. Its multi-layered graph structure and ability to efficiently navigate vector embeddings make it particularly appealing. However, despite its apparent advantages, HNSW may not be the optimal solution for large-scale and dynamic vector similarity search. In this blog post, we challenge the dominance of HNSW and explore why disk-based alternatives, such as IVF (Inverted File Index), might be more practical for massive datasets.
Pinecone integrates AI inferencing with vector database
(blocksandfiles.com)
GenAI inferencing can now be run directly from the Pinecone vector database to improve retrieval-augmented generation (RAG).
GenAI inferencing can now be run directly from the Pinecone vector database to improve retrieval-augmented generation (RAG).
Introducing integrated inference: Embed, rerank, and retrieve data with one API
(pinecone.io)
We’re excited to announce expanded inference capabilities alongside our core vector database to make it even easier and faster to build high-quality, knowledgeable AI applications with Pinecone.
We’re excited to announce expanded inference capabilities alongside our core vector database to make it even easier and faster to build high-quality, knowledgeable AI applications with Pinecone.
Scaling Document Data Extraction with LLMs and Vector Databases
(timescale.com)
Extracting structured data from unstructured documents is a powerful use case for large language models (LLMs). This sort of data extraction from complex documents has always remained a challenge. Done either completely manually or using current intelligent document processing (IDP) platforms that utilize previous-generation machine learning or natural language processing (NLP) techniques is very time-consuming and tedious.
Extracting structured data from unstructured documents is a powerful use case for large language models (LLMs). This sort of data extraction from complex documents has always remained a challenge. Done either completely manually or using current intelligent document processing (IDP) platforms that utilize previous-generation machine learning or natural language processing (NLP) techniques is very time-consuming and tedious.
The PlanetScale vectors public beta
(planetscale.com)
We're excited to announce that PlanetScale vector search and storage is now available in open beta! With PlanetScale vector support, you can store your vector data alongside your application's relational MySQL data — eliminating the need for a separate specialized vector database.
We're excited to announce that PlanetScale vector search and storage is now available in open beta! With PlanetScale vector support, you can store your vector data alongside your application's relational MySQL data — eliminating the need for a separate specialized vector database.
A Vector Database Plays Mario Kart 64
(medium.com)
In this article, I’ll introduce you to an original application of image search. I’ve named it Qdrant Kart, and, as you might guess, it involves using a Vector Database (Qdrant) to play Mario Kart 64 — one of my all-time favorite games.
In this article, I’ll introduce you to an original application of image search. I’ve named it Qdrant Kart, and, as you might guess, it involves using a Vector Database (Qdrant) to play Mario Kart 64 — one of my all-time favorite games.
BBQvec: An open-source, embedded vector index for Rust and Go
(daxe.ai)
At Daxe, we’re building Structured Semantic Search – a complete AI search stack. Our team leverages our collective experience from OpenAI, Google, Lyft, AWS, Harvard, Berkeley, and Darden to create novel technologies for developers and organizations to harness the full potential of their data.
At Daxe, we’re building Structured Semantic Search – a complete AI search stack. Our team leverages our collective experience from OpenAI, Google, Lyft, AWS, Harvard, Berkeley, and Darden to create novel technologies for developers and organizations to harness the full potential of their data.
Using the Pinecone vector database in .NET
(infoworld.com)
If you’re building generative AI applications, you need to control the data used to generate answers to user queries.
If you’re building generative AI applications, you need to control the data used to generate answers to user queries.
PGVector's Missing Features
(trieve.ai)
PGVector offers infrastructure simplicity at the cost of missing some key features desireable in search solutions. We explain what those are in this blog.
PGVector offers infrastructure simplicity at the cost of missing some key features desireable in search solutions. We explain what those are in this blog.
Create a RAG Pipeline with Pinecone
(vectorize.io)
This quickstart will walk you through creating and scheduling a pipeline that collects data from an Amazon S3 bucket, creates vector embeddings using an OpenAI embedding model, and writes the vectors to your Pinecone search index.
This quickstart will walk you through creating and scheduling a pipeline that collects data from an Amazon S3 bucket, creates vector embeddings using an OpenAI embedding model, and writes the vectors to your Pinecone search index.