---
license: apache-2.0
task_categories:
- robotics
tags:
- LeRobot
- so100
- test
- tsfile
- format:tsfile
- time-series
- timeseries
modality:
- tabular
- timeseries
configs:
- config_name: default
  data_files:
  - split: train
    path: data/pig2.tsfile
size_categories:
- 10K<n<100K
---

# pig2 (TsFile)

Apache TsFile version of [`pierfabre/pig2`](https://huggingface.co/datasets/pierfabre/pig2).

## Overview

This dataset is a LeRobot v2.1 dataset recorded with an SO-100 arm.

- **Robot:** SO-100 (so100)
- **Scale:** 50 episodes, 26,780 frames, 30 fps
- **Split:** train
- **Cameras (not uploaded):** `robot`, `webcam`

## Schema (TsFile structure)

The TsFile table is named `pig2`.

| Role | Columns |
| --- | --- |
| Time | `Time`, INT64 milliseconds |
| TAG | `episode_index`, `task_index` |
| FIELD | `frame_index`, `sample_index` |
| FIELD | `action_0` … `action_5`, FLOAT |
| FIELD | `observation_state_0` … `observation_state_5`, 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](https://huggingface.co/datasets/pierfabre/pig2/tree/main/videos).
- `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

```python
from tsfile import TsFileReader

with TsFileReader("data/pig2.tsfile") as reader:
    print(reader.get_all_table_schemas()["pig2"])
```

## Source & license

- Original dataset: https://huggingface.co/datasets/pierfabre/pig2
- Author: pierfabre
- License: apache-2.0