Episodes Preview Franka Panda Visualizer
4.51k episodes · 15 fps

DROID Visual Trajectories 5K TsFile

This dataset is a TsFile conversion of ASethi04/droid-trajectories-5k, the DROID Visual Trajectories dataset in LeRobot format. The source dataset contains visual object trajectories extracted from 4,506 DROID episodes using Gemini Robotics-ER and SAM2.

The robot data and referenced videos come from cadene/droid_1.0.1. Images and videos are not embedded here; the converted table preserves the droid_video_* path fields that point to the original DROID video files.

Dataset Details

  • Modalities: Time-series
  • Source robot type: panda
  • Split: train
  • Episodes: 4,506
  • Frames / TsFile rows: 943,315
  • Unique tasks: 1,429
  • Source chunks: 5
  • Sampling rate: 15 fps
  • Converted file: data/droid_trajectories_5k.tsfile
  • Table name: droid_trajectories_5k

TsFile Schema

Time is an INT64 millisecond timestamp synthesized as round(timestamp * 1000) from the source timestamp seconds column.

TAG columns:

  • episode_index
  • task_index

FIELD columns:

  • Episode/frame metadata: frame_index, sample_index, droid_episode_id
  • Task/video references: language_instruction, droid_video_exterior_1_left, droid_video_exterior_2_left, droid_video_wrist_left
  • Robot state/action: gripper_position, joint_position_0..6, actions_0..7
  • Visual trajectory: trajectory_center_x, trajectory_center_y, trajectory_xyxy_0..3, trajectory_labels

Conversion Notes

  • The source timestamp column is dropped because it is represented by the TsFile Time column in milliseconds.
  • The source index column is renamed to sample_index.
  • Fixed-width numeric vectors are flattened into scalar FLOAT fields: joint_position[7], actions[8], and trajectory_xyxy[4].
  • trajectory_labels is preserved as a JSON string because the source column is a variable-length list[string].
  • No numeric time-series rows are intentionally dropped.
  • Video files are not uploaded. Use the preserved droid_video_* fields with cadene/droid_1.0.1 to fetch the original videos.

Minimal Read Example

tsfile_path = "data/droid_trajectories_5k.tsfile"
table_name = "droid_trajectories_5k"

# Read the file with the Apache TsFile Java or Python SDK.
# Query a single episode with the episode_index TAG, for example:
# SELECT * FROM droid_trajectories_5k WHERE episode_index = 0

Source

Downloads last month
82