6 ms·
Turbovec – Google's TurboQuant for vector search in Rust
- esafak 2mo agolancedb and duckdb integrations would be great...
- nharada 2mo agoIt would be nice to have the README be a little more human written for a project where you actually want people to adopt it
- badatnames 2mo ago[flagged]
- ghm2199 2mo agoWow! 4GB for 10 million documents. This means one could build a reverse index much faster than before and devx processes like debugging, performance testing would become much smoother. Can't wait for the sqlite bindings to come out!
- ghm2199 2mo agoAlso the removal latency is on a log scale. Which is quite insane.
- burgerboii 2mo agoWho is this co-author called t <t@t>?
- cute_boi 2mo agoAs it is heavily vibe coded, I think member of technical staff at antropic has no clue.... Next Prompt: remove t@t and force commit.
- zuzululu 2mo agowhat could i use this for as part of my agentic workflow? codebase indexing? docs ?
- kyxsc 2mo agonotes/docs/wiki is a great use case
- spread2009 1mo ago[flagged]
- anishvarghese 2mo agoThis looks perfect for local, privacy first search, but since it's built in Rust, has anyone tried compiling it to WASM to run directly inside a browser extension?
- cpursley 2mo agoAlso interested.
- westurner 2mo agooxirs does embeddings and GraphRAG, and full text search with Tantivy; oxirs-vec, oxirs-graphrag There's an oxirs-wasm with RDF and SPARQL bindings with a query budget. Tantivy-wasm says that the release WASM bundle is 1.5 MB. cool-japan/oxirs: https://github.com/cool-japan/oxirs https://github.com/cool-japan/oxirs oxirs-wasm: https://crates.io/crates/oxirs-wasm https://crates.io/crates/oxirs-wasm tantivy-wasm: https://github.com/phiresky/tantivy-wasm https://github.com/phiresky/tantivy-wasm Is there an advantage to adding an MCP local memory interface over agent instructions on how to use a rust CLI? And then write Markdown documents with Google OKF-like frontmatter YAML metadata for agents that work with tokens not linked data graphs; https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md https://github.com/GoogleCloudPlatform/knowledge-catalog/blo...
- coredog64 2mo agoCan WASM use AVX512-VNNI?
- _tgxm 2mo agoIf anyone is looking to retrofit to an existing pipeline, I use similar ideas to compress vectors for job search, getting roughly 8x compression with about a 3.5% drop in quality. My experiment: https://corvi.careers/blog/vector-search-embedding-compression/ https://corvi.careers/blog/vector-search-embedding-compressi...
- spoaceman7777 2mo agoWell. That is insane. O_O Fantastic job!
- refulgentis 2mo agoBloviating nonsense, 3rd time I’ve seen something like this in HN since TurboQuant came out. You don’t need float32, never did. Source: I’ve been writing on device embedding code for 4 years.
- beernet 2mo agoWhy not just use Qdrant? They've been integrating TurboQuant for months, works well.
- kanungle 2mo agoIntegrated in 5 weeks and just expanded data types for turbo4 in last release. No longer need to store fp32 vectors if you don't need them
- cute_boi 2mo agoAnother vibe coded slop where they can't even spend time on Readme or documentation around code...
- Eridrus 2mo agoFAISS is no longer close to SoTA: https://ann-benchmarks.com/index.html https://ann-benchmarks.com/index.html https://vector-index-bench.github.io/ https://vector-index-bench.github.io/ https://big-ann-benchmarks.com/neurips23.html https://big-ann-benchmarks.com/neurips23.html
- nl 2mo agoI think their point is the size/performance tradeoff rather than outright performance. The point of TurboQuant is the size savings, while still giving high accuracy. It's been a while, but I do recall some high-performing vector matching indexes being very large.
- ehsanu1 2mo agoSurprised that usearch isn't in any of these, it's pretty fast.
- bobmarleybiceps 2mo agopeople should read turboquant's open review comments: https://openreview.net/forum?id=tO3ASKZlok https://openreview.net/forum?id=tO3ASKZlok
- esafak 2mo agotl,dr: there is an allegedly better alternative, and it's already implemented everywhere: https://github.com/VectorDB-NTU/RaBitQ-Library#rabitq-in-industry https://github.com/VectorDB-NTU/RaBitQ-Library#rabitq-in-ind...
- tracespect 2mo ago[flagged]
- cat-whisperer 2mo agoWhat's a good embedding model and search to run locally? something fast and lightweight.
- OutOfHere 2mo agoI am not convinced that Turbovec yields better retrieval than the same amount of bits of a Matryoshka embedding.
- myshapeprotocol 2mo ago[dead]
- lmeyerov 2mo agoInterestingly, while we don't fine-tune generative models for Louie.ai, we found fine-tuning embedding models to be a major $ saver. Instead of 1K-2K wide frontier embedding vector lens... Just 64. Huge savings on vector DB $$$. I'm curious how that works with something like turboquant. Not needed any more, still dominant, better together, ... .
- anthropic-dario 2mo ago[dead]
- mskkm 2mo agoThere are already several openreview comments alleging academic misconduct around TurboQuant: https://openreview.net/forum?id=tO3ASKZlok https://openreview.net/forum?id=tO3ASKZlok Some write-ups argue that this was deliberate rather than a good-faith mistake: https://dev.to/gaoj0017/turboquant-and-rabitq-what-the-public-story-gets-wrong-1i00 https://dev.to/gaoj0017/turboquant-and-rabitq-what-the-publi... And now this. Pretty bold AI slop.
- stlahxm 1mo ago[dead]
- emerthorn 1mo ago[dead]