Jon Bratseth
Jon Bratseth
All documents that _matches_ the query will be ranked by the first-phase ranking function, which can express any recency bias etc. as a mathematical expression, e.g. `( 0.75*bm25(title) + 0.25*bm25(body)...
> In this scenario, Vespa's weakAnd will still consider all such docs in first-phase ranking regardless of what k was set to. Is this accurate? Yes! > Is my understanding...
Remember to mark it @Beta
https://datasketches.apache.org/docs/Quantiles/QuantilesSketchOverview.html
No rush, but have you seen this @geirst ?
Take your time @geirst ... :-)
Thanks! Yes, more work is needed on this type resolving for sure. This is just the bare minimum to support both embed and pack_bits in the same statement.
You'd need to write your own [Embedder](https://docs.vespa.ai/en/reference/embedding-reference.html#custom-embedders). We don't have an example embedder using a remote endpoint, sorry.
You can use [global-phase](https://docs.vespa.ai/en/phased-ranking.html#using-a-global-phase-expression) instead of second-phase. With indexed search, second-phase is usually preferable because it runs locally in parallel on each content node, and as fan-out increases this becomes...
Vespa requires the network incl DNS to work before nodes are started so the usual issue is that people try to both start the containers and Vespa at the same...