# facebookresearch/MUSE

A library for Multilingual Unsupervised or Supervised word Embeddings

Repository: https://github.com/facebookresearch/MUSE
Canonical: https://ross.abutalabs.com/products/muse
Language: Python
License: NOASSERTION
License Family: other
Archived: true
Last push: 2022-08-31T10:12:54+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3236, "days_push": 1463, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3244, forks 558 (observed 2026-08-28T04:07:50.610842+00:00)

## What it is
MUSE is a Python library from Facebook AI Research for creating and aligning multilingual word embeddings, using both supervised (bilingual dictionary) and unsupervised (no parallel data) methods. It also provides fastText embeddings aligned in a common space, large-scale bilingual dictionaries, and evaluation datasets for cross-lingual word similarity and translation retrieval.

## Use cases
- align monolingual word embeddings from different languages into a shared space
- train unsupervised cross-lingual embeddings without parallel data
- evaluate word translation and cross-lingual word similarity
- build bilingual dictionaries for training and evaluation
- retrieve sentence translations using aligned embeddings
- run nearest neighbor search over multilingual embeddings with Faiss

## When to choose
- you need cross-lingual word embeddings or word translation without parallel corpora
- you want pre-aligned fastText embeddings and bilingual dictionaries for many language pairs
- you are doing NLP research on multilingual representation learning

## When to avoid
- you need modern contextual embeddings like multilingual BERT or sentence transformers
- you want a production-ready maintained library - the project is research code with infrequent updates
- you need sentence-level or document-level multilingual embeddings rather than word-level

## Facets
- artifact type: library
- maturity: maintenance
- function: nlp, machine-learning, search-engine
- domain: machine-learning
- platform: python
- tags: word-embeddings, multilingual-nlp, bilingual-dictionaries, cross-lingual, faiss, research, natural-language-processing, linux, gpu

## Member repositories
- facebookresearch/MUSE (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:50.610842+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 2026-08-29T18:43:54.835712+00:00, confidence not recorded.
  - readme: https://github.com/facebookresearch/MUSE (fetched 2026-08-28T04:07:50.610842+00:00, sha 7addd2d0c100)
- Data as of 2026-08-30T08:39:29.467469+00:00.