novastar1112025other

PushT PPO 1Hz Success Trajectories

PushT trajectory dataset generated from a trained 1Hz PPO policy in VisGym_3d, containing 200,000 successful episodes capped at 30 environment actions.

Downloads0
Episodes200,200

Why This Matters for Physical AI

This dataset provides large-scale vision-language trajectories from a trained reinforcement learning policy, enabling research into imitation learning and language-conditioned robotic manipulation in simulation.

Technical Profile

Modalities
rgblanguage
Action Space
end_effector_delta
Environment
simulation
Task Types
manipulationpush
Episodes
200,200
Data Format
JSONL
Annotation Types
language_instructionsreward_labelsaction_labels
License
other
Part of the VisGym family

Access

Need custom rgb data?

Claru builds purpose-built datasets for simulation applications with dense human annotations and quality assurance.

Request a Sample Pack

Related Datasets