Offline RL — learning policies from fixed datasets without further environment interaction — matters because most real-world domains (healthcare, robotics, autonomous driving) can't afford online exploration. CQL (2020) and IQL (2022) established strong baselines on the D4RL benchmark, but the field was disrupted by Decision Transformer (2021), which recast RL as sequence modeling. The latest wave uses pretrained language models as policy backbones, blurring the line between offline RL and in-context learning, with benchmarks like CORL tracking reproducibility across dozens of algorithms.
Canonical offline RL benchmark environment from D4RL. The halfcheetah-medium-v2 dataset contains 1M transitions collected from a medium-level SAC policy. Scores are reported as normalized return where 0 = random policy and 100 = expert SAC policy.
Leading models on D4RL HalfCheetah-Medium-v2.
No results yet. Be the first to contribute.
Didn't find the model, metric, or dataset you needed? Tell us in one line. We read every message and reply within 48 hours.
1 dataset tracked for this task.
Still looking for something on Offline RL? A missing model, a stale score, a benchmark we should cover — drop it here and we'll handle it.
Real humans read every message. We track what people are asking for and prioritize accordingly.