# naver/splade

SPLADE: sparse neural search (SIGIR21, SIGIR22)

Repository: https://github.com/naver/splade
Canonical: https://ross.abutalabs.com/products/splade
Language: Python
License: NOASSERTION
License Family: other
Topics: bert, information-retrieval, nlp, passage-retrieval, splade, sparse
Last push: 2024-05-03T14:52:32+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 1876, "days_push": 852, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1007, forks 96 (observed 2026-08-28T04:03:12.371649+00:00)

## What it is
SPLADE is a research library from NAVER for training, indexing, and retrieval with sparse neural search models based on BERT. It learns sparse query/document expansion representations that work with inverted indexes and includes evaluation on the BEIR benchmark.

## Use cases
- train a sparse neural retrieval model
- index documents with SPLADE for first-stage ranking
- evaluate neural retrievers on BEIR
- learn sparse query and document expansion with BERT
- build a passage retrieval system with inverted indexes
- reproduce SPLADE v2 research results

## When to choose
- you need interpretable sparse neural retrieval compatible with inverted indexes
- you want strong out-of-domain generalization on BEIR
- you are doing IR research on sparse expansion models

## When to avoid
- you need a production-ready plug-and-play search engine
- you want dense vector retrieval or hybrid search out of the box
- you need active community support or frequent updates

## Facets
- artifact type: library
- maturity: maintenance
- function: search-engine, machine-learning, nlp, rag
- domain: machine-learning
- platform: python
- tags: sparse-retrieval, bert, passage-ranking, inverted-index, beir-benchmark, neural-ir, search, natural-language-processing, information-retrieval, linux, gpu

## Member repositories
- naver/splade (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:12.371649+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:12:27.986845+00:00, confidence not recorded.
  - readme: https://github.com/naver/splade (fetched 2026-08-28T04:03:12.371649+00:00, sha 4dd43df5aacc)
- Data as of 2026-08-30T08:39:29.467469+00:00.