CoolFace
Datasetpublic

Sraghvi/exact-pusht-format-robot

Robot Dataset with EXACT Pusht Format Dataset with identical structure to working lerobot/pusht dataset. Format Details Structure: EXACT copy of lerobot/pusht file organization Files: chunk-000/file-000.parquet (not episode_000000.parquet) Episodes: Parquet format (not JSONL) Tasks: Parquet format (not JSONL) Features: Includes next.reward, next.done, next.success Files Structure data/chunk-000/file-000.parquet… See the full description on the dataset page: https://huggingface.co/datasets/Sraghvi/exact-pusht-format-robot.

sourceHugging Facemitupdated 1y agoView on Hugging Face
0likes27downloads
Dataset Card

Robot Dataset with EXACT Pusht Format

Dataset with identical structure to working lerobot/pusht dataset.

Format Details

  • —Structure: EXACT copy of lerobot/pusht file organization
  • —Files: chunk-000/file-000.parquet (not episode_000000.parquet)
  • —Episodes: Parquet format (not JSONL)
  • —Tasks: Parquet format (not JSONL)
  • —Features: Includes next.reward, next.done, next.success

Files Structure

data/chunk-000/file-000.parquet
videos/observation.image/chunk-000/file-000.mp4
meta/info.json
meta/stats.json
meta/tasks.parquet
meta/episodes/chunk-000/file-000.parquet

Data Features

  • —observation.state: 14-DOF robot state
  • —action: 14-DOF robot action
  • —next.reward: Task reward
  • —next.done: Episode termination
  • —next.success: Task success
  • —observation.image: Robot visualization video

Usage

Should work with LeRobot visualizer since it matches exact pusht structure.