Dataset Viewer
Auto-converted to Parquet Duplicate
The dataset viewer is not available for this split.
Parquet error: Scan size limit exceeded: attempted to read 317611167 bytes, limit is 300000000 bytes Make sure that 1. the Parquet files contain a page index to enable random access without loading entire row groups2. otherwise use smaller row-group sizes when serializing the Parquet files
Error code:   TooBigContentError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

tartanair-fish-v3-parquet

Per-frame Parquet dataset for ReCAST tracker training.

Schema

One row per frame, grouped by episode_id. Arrow memory-mapped access enables reading specific frames without loading entire episodes.

Column Type Description
episode_id int32 Episode identifier
frame_idx int32 Frame index within episode
jpeg binary JPEG-encoded RGB frame
depth list<float32> Flat H×W depth map
seg list<uint16> Semantic segmentation (empty if absent)
pose fixed_size_list<float64, 16> 4×4 camera-to-world pose (vision convention)
intrinsics fixed_size_list<float32, 4> [fx, fy, cx, cy] in pixels
gravity fixed_size_list<float32, 3> World-frame gravity direction
camera_model string "pinhole" or "equidistant"

Conventions

  • Poses: Vision convention (X=right, Y=down, Z=forward). Normalized during conversion.
  • Depth: Projective z-depth (all datasets, normalized during conversion).
  • Gravity: Unit vector pointing downward in world frame.

Provenance

  • Dataset: tartanair
  • Episodes: 0
  • Frames: 0
  • Created: 2026-03-15T19:46:29.610714+00:00
  • Build script: scripts/convert_episodes_to_parquet.py
  • Git commit: 652912ed324118c3d69a652be186f0ca129de190
Downloads last month
26