3 ms·
> - multiple vectors per document Can I have (multiple) vectors for a single field? That would be quite helpful.
by freilanzer 3y ago
> - multiple vectors per document
Can I have (multiple) vectors for a single field? That would be quite helpful.
- jkb79 3y agoYes, Vespa has a generic Tensor framework that allows you to index multiple vectors for a single field, see https://blog.vespa.ai/semantic-search-with-multi-vector-indexing/ https://blog.vespa.ai/semantic-search-with-multi-vector-inde... for details. field embeddings type tensor<float>(p{}, x[384]) to represent a multi-vector field { "0": [0.1....], "1": [0.2,..] }
- dathinab 3y agoyes that is what I meant generally if you have multiple embeddings for the same document you have two choices: - create one document for each embedding and make sure non membedding specific attributes are the same across all of this document clones -- vespa makes this more convenient by having child documents - have a field with multiple documents, i.e. there are multipel vectors in the HNSW-index which point to the same document -- vespa support this, too. It's what I meant. vespa is currently the only vector search enabled search system which supports both in a convenient way, but then there are so many "vector databases" poping up every month that I might have missed some