THULab/snow_sb

SNOW.SB (TsFile)

Apache TsFile version of williamdgomez/SNOW.SB.

Overview

This dataset is a LeRobot v3.0 dataset recorded with a Lekiwi client robot at 24 fps.

• Robot: Lekiwi client (lekiwi_client)
• Scale: 91 episodes, 73,578 frames, 24 fps
• Split: train
• Cameras (not uploaded): front, wrist, top

Schema (TsFile structure)

The TsFile table is named snow_sb.

Role	Columns
Time	Time, INT64 milliseconds
TAG	episode_index, task_index
FIELD	frame_index, sample_index
FIELD	action_0 … action_8
FIELD	observation_state_0 … observation_state_8

Vector columns are flattened into scalar FLOAT measurements (single precision). The source index column is retained as sample_index.

Conversion notes

• Time = round(timestamp * 1000) with millisecond precision; time restarts inside each episode, while episode_index and task_index identify the TsFile device.
• The original timestamp field is omitted because it is exactly represented by Time / 1000.
• Camera video streams are NOT included in this repository; they remain in the source videos tree.
• meta/ is mirrored from the source. Aside from the redundant timestamp column and the excluded videos, no source rows or numeric fields are dropped.

Read example

from tsfile import TsFileReader

with TsFileReader("data/snow_sb.tsfile") as reader:
    print(reader.get_all_table_schemas()["snow_sb"])

Source & license

• Original dataset: https://huggingface.co/datasets/williamdgomez/SNOW.SB
• Author: williamdgomez
• License: apache-2.0