Train Offline, Test Online: A Real Robot Learning Benchmark
This benchmark addresses accessibility and standardization issues in robot learning research, enabling broader participation and direct method comparisons without hardware costs, though it is incremental in building on existing benchmark concepts.
The authors tackled the challenges of expensive robots, lack of generalization across labs, and insufficient robotics data by introducing the Train Offline, Test Online (TOTO) benchmark, which provides remote access to shared hardware and an open-source dataset for offline training, with initial results comparing five visual representations and four policy learning baselines.
Three challenges limit the progress of robot learning research: robots are expensive (few labs can participate), everyone uses different robots (findings do not generalize across labs), and we lack internet-scale robotics data. We take on these challenges via a new benchmark: Train Offline, Test Online (TOTO). TOTO provides remote users with access to shared robotic hardware for evaluating methods on common tasks and an open-source dataset of these tasks for offline training. Its manipulation task suite requires challenging generalization to unseen objects, positions, and lighting. We present initial results on TOTO comparing five pretrained visual representations and four offline policy learning baselines, remotely contributed by five institutions. The real promise of TOTO, however, lies in the future: we release the benchmark for additional submissions from any user, enabling easy, direct comparison to several methods without the need to obtain hardware or collect data.