Getting Started

edit page

• Mac / Linux
• Windows
• Vagrant

Moving from the minimal quick start to more advanced use cases

• Tutorial: Blog Search. Vespa tutorial introducing read and write API's, Vespa's query language, YQL, ranking and relevancy, faceted search (grouping) and sorting.
• Tutorial: Text Search. A text search tutorial and introduction to text ranking with Vespa using traditional information retrieval techniques like BM25.
• Tutorial: Improving Text Search with Machine Learning. This tutorial builds on the text search tutorial but introduces Learning to Rank to improve relevancy.
• Tutorial: Text Search using Vectors. This tutorial introduce using pre-trained embedding models to do text search. It uses Vespa support for nearest neighbor search, there is also support for fast approximate nearest neighbor search in Vespa. Also see Semantic Retrieval for Question Answering Applications, with examples for doing hybrid retrieval using both sparse (term based) and dense (vector) based search and retrieval.

Vespa is great for search use cases, but also supports advanced real time online serving of recommender models - see how in these tutorials:

• Tutorial: Blog Recommendation. This builds on the Blog Search tutorial and uses collaborative filtering to recommend blog posts.
• Tutorial: Blog recommendation with Neural Network models.

From 10 blue links to answering questions. Try the Question Answering use case. This use case demonstrates several of the advanced capabilities of Vespa. Ranking with Transformer (e.g. BERT) models, dense retrieval using vectors, multi-phased retrieval and ranking, and custom plugins to build an end-to-end application.

The e-commerce shopping use case demonstrates Vespa grouping, true in-place partial updates, custom ranking and more.

Vespa is an open source platform and can be deployed in multiple ways. These guides are good starters for building multi-node clusters:

• quick-start-kubernetes (minikube) and basic-search-on-gke
• quick-start-multinode-aws
• quick-start-multinode-aws-ecs
• basic-search-on-docker-swarm
• Monitoring Vespa with Grafana

Vespa supports writing plugins in languages that runs in the JVM. Use this to simplify query and document processing and build serving logic close to the content nodes, avoid network hops and overhead of HTTP(s):

• album-recommendation-searcher for Searcher development
• album-recommendation-docproc for Document Processor development