# google/uis-rnn

This is the library for the Unbounded Interleaved-State Recurrent Neural Network (UIS-RNN) algorithm, corresponding to the paper Fully Supervised Speaker Diarization.

Repository: https://github.com/google/uis-rnn
Canonical: https://ross.abutalabs.com/products/uis-rnn
Homepage: https://arxiv.org/abs/1810.04719
Language: Python
License: Apache-2.0
License Family: permissive
Topics: speaker-diarization, uis-rnn, speaker-recognition, supervised-learning, clustering, supervised-clustering, machine-learning
Archived: true
Last push: 2024-09-25T14:33:18+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2871, "days_push": 707, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1588, forks 320 (observed 2026-08-28T04:05:08.032197+00:00)

## What it is
A Python library implementing the Unbounded Interleaved-State Recurrent Neural Network (UIS-RNN) algorithm for segmenting and clustering sequential data, originally published in the paper 'Fully Supervised Speaker Diarization'. It models each speaker with a parameter-sharing RNN integrated with a distance-dependent Chinese Restaurant Process to handle an unknown number of speakers.

## Use cases
- diarize who spoke when in a multi-speaker audio recording
- cluster speaker embeddings into individual speakers
- segment and label sequential data with an unknown number of classes
- train a supervised clustering model from labeled speaker examples
- perform online speaker diarization on d-vector embeddings

## When to choose
- you have speaker embeddings (e.g., d-vectors) and need to assign them to speakers
- you need online/real-time diarization rather than offline spectral clustering
- the number of speakers is unknown in advance
- you have labeled training data for supervised clustering

## When to avoid
- you need a complete end-to-end diarization pipeline including speech detection and embedding extraction
- you want an actively developed tool with frequent updates
- you need offline-only clustering and prefer simpler methods like spectral clustering
- you cannot provide labeled training data

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, audio-processing, speech-recognition
- domain: machine-learning, speech-processing
- platform: python, cross-platform
- tags: speaker-diarization, uis-rnn, clustering, pytorch, supervised-learning, sequence-modeling, audio

## Member repositories
- google/uis-rnn (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:08.032197+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 2026-08-30T03:55:01.049026+00:00, confidence not recorded.
  - readme: https://github.com/google/uis-rnn (fetched 2026-08-28T04:05:08.032197+00:00, sha 11348c25c57b)
  - homepage: https://arxiv.org/abs/1810.04719 (fetched 2026-08-29T11:25:48.604599+00:00, sha 2c7ebcb7c061)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T11:25:48.613844+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T11:25:48.617491+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T11:25:48.620319+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T11:25:48.615672+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.