All Tools
L
AgentsFreeOpen Source
LIBERO
Lifelong robot learning benchmark for knowledge transfer
MIT
ABOUT
Robot agents forget old skills when fine-tuned on new tasks. LIBERO benchmarks knowledge transfer across language-conditioned manipulation suites so researchers measure lifelong learning instead of single-task success.
INTEGRATION GUIDE
1. Benchmark lifelong manipulation agents on LIBERO task suites
2. Train language-conditioned policies that transfer across object sets
3. Evaluate catastrophic forgetting in sequential robot learning
4. Collect demonstration data for imitation on LIBERO environments
TAGS
pythonagentsroboticslifelong-learningmanipulationbenchmarkimitation-learning