ann-benchmarks VS txtai

On top of my head, pgvector only supports 2 indexes, those are running in memory only. They don't support GPU indexing, nor Disk based indexing, they also don't have separation of query and insertions.

Also with different people I've talked to, they struggle with scale past 100K-1M vector.

You can also have a look yourself from a performance perspective: https://ann-benchmarks.com/

https://ann-benchmarks.com/ is a good resource covering those libraries and much more.

We utilized the ANN Benchmarks methodology, a standard for benchmarking vector databases. Our tests used the dbpedia dataset of 1,000,000 OpenAI embeddings (1536 dimensions) and inner product distance metric for both Pinecone and pgvector.

Data warehouses are columnar stores. They are very different from row-oriented databases - like Postgres, MySQL. Operations on columns - e.g., aggregations (mean of a column) are very efficient.

Most vector databases use one of a few different vector indexing libraries - FAISS, hnswlib, and scann (google only) are popular. The newer vector dbs, like weaviate, have introduced their own indexes, but i haven't seen any performance difference -

Reference: https://ann-benchmarks.com/

(Blog author here). Thanks for the question. In this case the index for both DiskANN and pgvector HNSW is small enough to fit in memory on the machine (8GB RAM), so there's no need to touch the SSD. We plan to test on a config where the index size is larger than memory (we couldn't this time due to limitations in ANN benchmarks [0], the tool we use).

To your question about RAM usage, we provide a graph of index size. When enabling PQ, our new index is 10x smaller than pgvector HNSW. We don't have numbers for HNSWPQ in FAISS yet.

[0]: https://github.com/erikbern/ann-benchmarks/

In terms of "All You Need" for Vector Search, ANN Benchmarks (https://ann-benchmarks.com/) is a good site to review when deciding what you need. As with anything complex, there often isn't a universal solution.

txtai (https://github.com/neuml/txtai) can build indexes with Faiss, Hnswlib and Annoy. All 3 libraries have been around at least 4 years and are mature. txtai also supports storing metadata in SQLite, DuckDB and the next release will support any JSON-capable database supported by SQLAlchemy (Postgres, MariaDB/MySQL, etc).

pg_vector doesn't perform well compared to other methods, at least according to ANN-Benchmarks (https://ann-benchmarks.com/).

txtai is more than just a vector database. It also has a built-in graph component for topic modeling that utilizes the vector index to autogenerate relationships. It can store metadata in SQLite/DuckDB with support for other databases coming. It has support for running LLM prompts right with the data, similar to a stored procedure, through workflows. And it has built-in support for vectorizing data into vectors.

For vector databases that simply store vectors, I agree that it's nothing more than just a different index type.

The reason https://ann-benchmarks.com is so good, is that we can see a plot of recall vs latency. I can see you have some latency numbers in the leaderboard at the bottom, but it's very difficult to make a decision.

As a practitioner that works with vector databases every day, just latency is meaningless to me, because I need to know if it's fast AND accurate, and what the tradeoff is! You can't have it both ways. So it would be helpful if you showed plots showing this tradeoff, similar to ann-benchmarks.

I tend to agree with this sentiment. Many junior devs and/or those in college want to contribute. Then they feel entitled to merge a PR that they worked hard on often without guidance. I'm all for working with people but projects have standards and not all ideas make sense. In many cases, especially with commercial open source, the project is the base of a companies identity. So it's not just for drive-by ideas to pad a resume or finish a school project.

For those who do want to do this, I'd recommend writing an issue and/or reaching out to the developers to engage in a dialogue. This takes work but it will increase the likelihood of a PR being merged.

Disclaimer: I'm the primary developer of txtai (https://github.com/neuml/txtai), an open-source vector database + RAG framework

txtai is an all-in-one embeddings database for semantic search, LLM orchestration and language model workflows.

Bootstrapping only works if you have the runway to do it and you don't feel the need to grow fast.

With NeuML (https://neuml.com), I've went the bootstrapping route. I've been able to build a fairly successful open source project (txtai 6K stars https://github.com/neuml/txtai) and a revenue positive company. It's a "live within your means" strategy.

VC funding can have a snowball effect where you need more and more. Then you're in the loop of needing funding rounds to survive. The hope is someday you're acquired or start turning a profit.

I would say both have their pros and cons. Not all ideas have the luxury of time.

I agree that in many cases people are puffing their feathers to try to be something they're not (at least not yet). Some believe in the fake it until you make it mentality.

With NeuML (https://neuml.com), the website is a simple HTML page. On social media, I'm honest about what NeuML is, that I'm in my 40s with a family and not striving to be the next Steve Jobs. I've been able to build a fairly successful open source project (txtai 6K stars https://github.com/neuml/txtai) and a revenue positive company. For me, authenticity and being genuine is most important. I would say that being genuine has been way more of an asset than liability.

I'll add txtai (https://github.com/neuml/txtai) to the list.

There is still plenty of room for innovation in this space. Just need to focus on the right projects that are innovating and not the ones (re)working on problems solved in 2020/2021.

txtai is an all-in-one embeddings database for semantic search, LLM orchestration and language model workflows.

Nice project! I've long used Tika for document parsing given it's maturity and wide number of formats supported. The XHTML output helps with chunking documents for RAG.

Here's a couple examples:

- https://neuml.hashnode.dev/build-rag-pipelines-with-txtai

- https://neuml.hashnode.dev/extract-text-from-documents

Disclaimer: I'm the primary author of txtai (https://github.com/neuml/txtai).

If you're interested in graphs + RAG and want an alternate approach, txtai has a semantic graph component.

https://neuml.hashnode.dev/introducing-the-semantic-graph

https://github.com/neuml/txtai

Disclaimer: I'm the primary author of txtai