vespa.ai
@vespaengine
Followers
4K
Following
828
Media
35
Statuses
538
https://t.co/abkb8IjPSH - the open source platform for combining data and AI, online. Vectors/tensors, full-text, structured data; ML model inference at scale.
Joined September 2017
In our latest episode of the Vespa Voice podcast with Ravindra Harige, founder of Searchplex, we explore a common but often misdiagnosed issue in modern software: search problems hiding in plain sight. https://t.co/bU2ihuzehx
1
0
6
This hack isn't for everyone, but in some situations it's possible and useful to turn off summary fetching.
0
0
5
If you want to create your own RAG application with this level of quality, clone the open source RAG Blueprint. https://t.co/GwlxnOlVs1
github.com
Repository of sample applications for https://vespa.ai, the open big data serving engine - vespa-engine/sample-apps
Two great alternatives, both built on
1
2
8
Filtered vector search is a massively important and overlooked problem for RAG and vector DBs. Very excited to see this new blog post from @vespaengine detailing its implementation of ACORN, along with many clever extensions to deliver huge speedups for search with filters.
In real vector search systems, performance is dominated by combining it efficiently with filters. Few test this properly. 🧵
1
8
24
Two great alternatives, both built on
0
2
9
Built on
Perplexity Search API: Providing direct search results in milliseconds for grounding LLMs and agents with real-time information from the web. This is an effort that began more than two years ago: to build our own search index. So much progress in a short period of time. We look
3
5
35
Vector search alone isn’t enough. Production-grade AI search needs more: combining semantic, keyword and metadata retrieval, applying machine-learned ranking and handling constantly changing structured and unstructured data, all at scale. Thanks to @vespaengine
thenewstack.io
Users expect search not just to return accurate results, but to do the heavy lifting: Answer a question, summarize research, or even solve a problem.
0
2
5
Plus lots of new helpful content. Read it here:
blog.vespa.ai
Advances in Vespa features and performance include new ANN tuning parameters, improvements to Geo filtering, filtering in grouping, and relevance score carry-over to global ranking.
0
0
2
Our August newsletter is out! New feature highlights: - New ANN optimizations that lets you optimize recall+cost when combining vector search and filters. - Binary data detection to protect against bogus writes creating havoc. - Filtering in the grouping language. - Geo
1
2
12
You can also read a full explanation of what these parameters do here:
blog.vespa.ai
This blog post highlights the latest additions to HNSW in Vespa, how to use them, and what’s to come in the future.
0
2
7
We'll soon update defaults to give everybody improved performance with no effort, but to really get the best performance you should tune to your case. We have made a guide for that:
blog.vespa.ai
This a companion post to the previous technical blog post, explaining how to tweak Vespa’s ANN parameters.
1
1
7
To optimize all parts of the filter space (make all queries efficient), you need to combine different strategies
1
2
3
We have introduced new tuning parameters in Vespa that lets you improve recall and cost, inspired by Acorn and beam search papers. They *really* help.
1
1
4
What you need to test is performance AND recall at various filter strengts. The challenging area is around 80-99% filter strength, and the devil has made it so that this is where most real-world applications live.
1
1
3
In real vector search systems, performance is dominated by combining it efficiently with filters. Few test this properly. 🧵
1
5
20
ACORN-1 and Adaptive Beam Search have been in @vespaengine for a while, but now we have a detailed post about how it works: https://t.co/gUF9vh330j
1
2
10
We made a new video to explain the point of Vespa's architecture
0
3
9
We just added a guide to creating, embedding, retrieving, ranking and selecting chunks. Probably contains some things you didn't know.
1
0
6
We just did a podcast about the process of migration (trade-offs included) from #Elasticsearch to @vespaengine With @dainius_jocas and @KevinPetrieTech 🙌 https://t.co/BtCGVyzP2U
em360tech.com
In this episode of the Don't Panic, It's Just Data podcast, Kevin Petrie, VP of Research at BARC and the podcast host, is joined by Dainius Jocas, Search Engineer at Vinted, and Radu Gheorghe,...
0
4
6