Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning

Open paper
Benchmark Model Rank Results
meta-learning-on-ml10MAML#1Meta-test success rate: 36%Meta-train success rate: 25%
meta-learning-on-ml10RL^2#2Meta-test success rate: 10%Meta-train success rate: 50%
meta-learning-on-ml10PEARL#3Meta-test success rate: 0%Meta-train success rate: 42.78%
meta-learning-on-mt50Multi-task multi-head SAC#2Average Success Rate: 35.85%

Rank counts only results with a code link.