# RecBole

A unified, comprehensive and efficient recommendation library

Repository: https://github.com/RUCAIBox/RecBole
Canonical: https://ross.abutalabs.com/products/recbole
Homepage: https://recbole.io/
Language: Python
License: MIT
License Family: permissive
Topics: recommender-systems, collaborative-filtering, knowledge-graph, ctr-prediction, deep-learning, pytorch, graph-neural-networks, sequential-recommendation, recommendations, recommendation-system, recommender
Last push: 2025-02-24T07:16:41+00:00
Link (homepage): https://recbole.io/

## Health v2 (maintenance only)
Score: 26/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 8, release rhythm 8, longevity 100
- inputs: {"age_days": 2274, "days_push": 555, "days_rel": 556, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4541, forks 747 (observed 2026-08-28T04:08:53.121148+00:00)

## What it is
RecBole is a unified, comprehensive and efficient recommendation library built on Python and PyTorch for reproducing and developing recommendation algorithms. It implements 94 algorithms across general, sequential, context-aware, and knowledge-based recommendation, with support for 44 benchmark datasets and standardized evaluation protocols.

## Use cases
- reproduce and benchmark recommendation algorithms
- train collaborative filtering models on user-item interaction data
- run sequential recommendation experiments with transformers and RNNs
- evaluate context-aware and CTR prediction models
- experiment with knowledge-graph-based recommenders
- compare recommenders under standard evaluation protocols

## When to choose
- you need a unified PyTorch framework for recommender systems research
- you want ready implementations of many classic and deep learning recommendation models
- you need standardized datasets and evaluation protocols for fair comparison
- you want GPU-accelerated training of recommendation models

## When to avoid
- you need a production recommendation serving system rather than a research library
- your project is not based on Python and PyTorch
- you need a simple plug-and-play recommender with minimal configuration
- you require a managed or hosted recommendation service

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, deep-learning, data-science, benchmarking
- domain: machine-learning, deep-learning, data-science
- platform: python, cross-platform
- tags: recommender-systems, collaborative-filtering, ctr-prediction, sequential-recommendation, knowledge-graph, pytorch, graph-neural-networks, research, algorithms, gpu

## Member repositories
- RUCAIBox/RecBole (main) score 26
- RUCAIBox/RecSysDatasets (docs) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:53.121148+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 2026-08-29T18:20:09.254575+00:00, confidence not recorded.
  - readme: https://github.com/RUCAIBox/RecBole (fetched 2026-08-28T04:08:53.121148+00:00, sha 41659e61736f)
  - homepage: https://recbole.io/ (fetched 2026-08-29T09:05:54.257702+00:00, sha 37a69404ea93)
  - registry_pypi: https://pypi.org/pypi/recbole/json (fetched 2026-08-29T09:05:54.269913+00:00, sha 099027edd46a)
- Data as of 2026-08-30T08:39:29.467469+00:00.