pusht / README.md
smilegeng's picture
Upload README.md with huggingface_hub
3605b9f verified
|
Raw
History Blame Contribute Delete
3.17 kB
---
license: mit
task_categories:
- robotics
tags:
- LeRobot
- tsfile
- time-series
- timeseries
---
# lerobot_pusht (TsFile 格式)
> 本数据集由 HuggingFace 数据集 [**lerobot/pusht**](https://huggingface.co/datasets/lerobot/pusht) 转换为 [Apache TsFile](https://tsfile.apache.org/) 格式。
> **原始数据集**:https://huggingface.co/datasets/lerobot/pusht
## 数据集简介
Push-T 是 **Diffusion Policy**(Chi et al. 2023)提出的机械臂操作基准任务:智能体需把一个 T 形物块推到目标位姿。本数据集使用 [LeRobot](https://github.com/huggingface/lerobot) 采集(codebase v2.0),是其**低维状态版本**——二维末端执行器状态与二维动作的时间序列。
- **Homepage**:https://diffusion-policy.cs.columbia.edu/
- **Paper**:https://arxiv.org/abs/2303.04137v5
- **License**:MIT
- **规模**:206 episodes / 25,650 frames / 1 task
- **采样率**:10 fps(每帧 100 ms)
## 原始数据结构
| 列 | 类型 | 说明 |
|----|------|------|
| `observation.state` | float32[2] | 末端执行器状态(x, y) |
| `action` | float32[2] | 动作(x, y) |
| `episode_index` | int64 | episode 序号 |
| `frame_index` | int64 | episode 内帧序号 |
| `timestamp` | float32 | episode 内相对秒数 |
| `next.reward` | float32 | 奖励 |
| `next.done` / `next.success` | bool | 终止 / 成功标志 |
| `index` | int64 | 全局样本序号 |
| `task_index` | int64 | 任务序号 |
| `observation.image` | video[96×96×3] | (**未包含**,本低维版本不含图像) |
## 转换为 TsFile 的说明
转换走「脚本预处理 + Apache TsFile Java 工具(schema 模式)」路径:
- **数组展开**`observation.state[2]``state_0``state_1``action[2]``action_0``action_1`(保持 float32 单精度 → TsFile `FLOAT`)。
- **列名清洗**`.` 统一替换为 `_``next.reward``next_reward` 等)。
- **时间轴**`Time = frame_index × 100ms`(10 fps),时间精度 `ms`
- **Tag 列(设备维度)**`episode_id``task_id` 声明为 TsFile **TAG**,因此每个 episode 是独立 device,各自时间轴从 0 开始、互不冲突。
- **⚠️ 丢弃的列**:原始 `timestamp`(episode 相对秒数,float)被丢弃,因为它在不同 episode 间会重复;改用 `frame_index × 100ms` 的整数毫秒时间轴。其余列全部保留。
转换后表名为 `pusht`,单文件 `lerobot_pusht.tsfile`,共 25,650 行。
## 用法
```python
# 用 Apache TsFile SDK 读取 lerobot_pusht.tsfile
from tsfile import TsFileReader
reader = TsFileReader("lerobot_pusht.tsfile")
# table "pusht":tag 列 episode_id / task_id,其余为 field 测点
```
## 引用
```bibtex
@article{chi2024diffusionpolicy,
author = {Cheng Chi and Zhenjia Xu and Siyuan Feng and Eric Cousineau and Yilun Du and Benjamin Burchfiel and Russ Tedrake and Shuran Song},
title ={Diffusion Policy: Visuomotor Policy Learning via Action Diffusion},
journal = {The International Journal of Robotics Research},
year = {2024},
}
```