chomeed/mimicgen_three_piece_assembly_d0_224x224
mimicgen_three_piece_assembly_d0_224x224 MimicGen ThreePieceAssembly_D0 demonstrations in LeRobot v3 format, with both cameras rendered at 224x224. Same pipeline and conventions as chomeed/mimicgen_coffee_d1_224x224. Note this task's demonstrations are not uniformly successful: 18/1000 never reach the full reward, and 298/1000 reach it and then lose it before the final frame, so success genuinely varies within and across episodes. episodes 1000 frames 336695 fps… See the full description on the dataset page: https://huggingface.co/datasets/chomeed/mimicgen_three_piece_assembly_d0_224x224.
mimicgenthreepieceassemblyd0_224x224
MimicGen ThreePieceAssembly_D0 demonstrations in LeRobot v3 format, with both cameras rendered at 224x224. Same pipeline and conventions as chomeed/mimicgen_coffee_d1_224x224. Note this task's demonstrations are not uniformly successful: 18/1000 never reach the full reward, and 298/1000 reach it and then lose it before the final frame, so success genuinely varies within and across episodes.
Provenance
Built from NVIDIA's released core/three_piece_assembly_d0.hdf5 (amandlek/mimicgen_datasets). That file already contains every low-dim observation, but its images are only 84x84, so only the cameras were recomputed: each frame's recorded states vector is written back into a live ThreePieceAssembly_D0 and both views are re-rendered at 224x224. Every other column is copied verbatim from the source hdf5, so this is a re-render, not a re-simulation, and no dynamics were replayed.
Fidelity check: re-rendering demo 0 at the source's own 84x84 and diffing against its stored images gives mean |difference| of FIDELITY_3P. The small residual is expected -- the originals were rendered on a different OpenGL stack.
Schema
Follows the conventions of chomeed/mimicgen_coffee_d0_224x224:
observation.state(51) =concat(robot0_eef_pos, robot0_eef_quat, robot0_gripper_qpos, object), i.e. 9 proprioceptive dims followed by 42 privileged object dims. Slicing to the leading 9 gives a proprioception-only policy input.observation.sim_state(58) replays into a liveThreePieceAssembly_D0.success == done == (reward >= 1.0).source_demo_index == episode_index-- the demo's index in the source hdf5.is_first/is_last/is_terminalmark only the first and last frame of an episode.
Column widths are task-specific: coffee carries a 57-dim object block plus *_rel_pod keys, square a 17-dim block with no rel_* keys.
The robot is read from the source env_args: Panda.
License
cc-by-4.0, inherited from the MimicGen datasets this is derived from.