Check your teleop data before you train
Step through a few episodes in the LeRobot dataset visualizer before spending GPU hours.
Look for dropped or frozen camera frames, cameras out of sync with joint states, long idle stretches at the start and end, and task descriptions that do not match what happens.
Referenced repos
Indexed from Hugging Face and linked upstream, not hosted here.
A browser tool for stepping through LeRobot-format datasets on the Hub, showing each episode's camera streams next to plots of joint states and actions.
About 200 human demonstrations in a 2D simulation where a round pusher slides a T-shaped block onto a target outline, stored in LeRobot v3 format.
A filtered collection of 128 SO-100 datasets from 55 community contributors (about 11k episodes and 5.1M frames), with hand-cleaned task descriptions, released as the pretraining subset for SmolVLA.