# Lifelong-Robot-Learning/LIBERO

Benchmarking Knowledge Transfer in Lifelong Robot Learning

Repository: https://github.com/Lifelong-Robot-Learning/LIBERO
Canonical: https://ross.abutalabs.com/products/libero
Language: Jupyter Notebook
License: MIT
License Family: permissive
Topics: benchmark, imitation-learning, lifelong-learning, manipulation, multitask-learning, robot-learning
Last push: 2025-03-15T12:14:04+00:00

## Health v2 (maintenance only)
Score: 35/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 11, release rhythm 35, longevity 88
- inputs: {"age_days": 1243, "days_push": 536, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2238, forks 462 (observed 2026-08-28T04:06:29.548017+00:00)

## What it is
LIBERO is a benchmark for studying knowledge transfer in multitask and lifelong robot learning, built on a procedural generation pipeline for manipulation tasks. It provides 130 tasks across four suites, datasets, policy architectures, and lifelong learning algorithm baselines.

## Use cases
- benchmark knowledge transfer in lifelong robot learning
- evaluate imitation learning policies on manipulation tasks
- study multitask learning with controlled distribution shifts
- pretrain and test policies on LIBERO-90/LIBERO-10 splits
- compare lifelong learning algorithms against sequential finetuning baselines

## When to choose
- you need standardized manipulation task suites for continual/lifelong learning research
- you want procedurally generated robot manipulation tasks with datasets and baselines

## When to avoid
- you need real-robot hardware experiments rather than simulation benchmarks
- you work outside robot learning, e.g. NLP or vision-only benchmarks

## Facets
- artifact type: dataset
- maturity: active
- function: benchmarking, machine-learning, simulation, testing
- domain: robotics, machine-learning, reinforcement-learning, simulation
- platform: python
- tags: imitation-learning, lifelong-learning, manipulation, multitask-learning, robot-learning, benchmark, procedural-generation, continual-learning, linux, gpu

## Member repositories
- Lifelong-Robot-Learning/LIBERO (main) score 35

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:29.548017+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 2026-08-30T02:44:12.592497+00:00, confidence not recorded.
  - readme: https://github.com/Lifelong-Robot-Learning/LIBERO (fetched 2026-08-28T04:06:29.548017+00:00, sha 033c717d514a)
- Data as of 2026-08-30T08:39:29.467469+00:00.