THULab/base4_plate_croissant_02_BC_FV

base4-plate-croissant-02-BC-FV (TsFile)

Apache TsFile version of maskjp/base4-plate-croissant-02-BC-FV.

Overview

This dataset is a LeRobot v2.1 dataset of a U850 arm placing a croissant on a plate (BC-FV recording), with base pose, state and action vectors.

• Robot: U850 (u850)
• Scale: 50 episodes, 73,213 frames, 50 fps
• Split: train
• Cameras (not uploaded): left, right, wrist

Schema (TsFile structure)

The TsFile table is named base4_plate_croissant_02_BC_FV.

Role	Columns
Time	Time, INT64 milliseconds
TAG	episode_index, task_index
FIELD	frame_index, sample_index, next_done, correction
FIELD	observation_state_0 … observation_state_9, FLOAT
FIELD	observation_base_pose_0 … observation_base_pose_2, FLOAT
FIELD	action_0 … action_9, FLOAT

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/base4_plate_croissant_02_BC_FV.tsfile") as reader:
    print(reader.get_all_table_schemas()["base4_plate_croissant_02_BC_FV"])

Source & license

• Original dataset: https://huggingface.co/datasets/maskjp/base4-plate-croissant-02-BC-FV
• Author: maskjp
• License: apache-2.0