Joocjun/GR1-Tabletop-NextState-1000x24

---
license: apache-2.0
task_categories:
- robotics
tags:
- LeRobot
- GR00T
- GR1
- robocasa
- simulation
- manipulation
size_categories:
- 10K<n<100K
---

# GR1 Tabletop Merged LeRobot Datasets

Merged and subsampled versions of the GR1 tabletop manipulation datasets from the [NVIDIA PhysicalAI-Robotics-GR00T-X-Embodiment-Sim](https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim) collection, formatted in **LeRobot v2.0** format.

## Dataset Variants

| Variant | Demos/Task | Tasks | Total Episodes | Total Frames | Approx Size |
|---------|-----------|-------|----------------|--------------|-------------|
| `1000x24/` | 1000 | 24 folders, 186 unique tasks | 24,000 | 6,020,058 | ~40 GB |
| `300x24/` | 300 | 24 folders, 186 unique tasks | 7,200 | 1,803,236 | ~12 GB |
| `100x24/` | 100 | 24 folders, 185 unique tasks | 2,400 | 602,846 | ~4 GB |

## Source Dataset

The original per-task datasets were downloaded from:
**[nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim](https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim)**

To download the original dataset:
```bash
# Install huggingface_hub
pip install huggingface_hub

# Download specific task datasets
from huggingface_hub import snapshot_download
snapshot_download(
repo_id="nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim",
repo_type="dataset",
allow_patterns="gr1_unified.PnP*", # or specific task patterns
local_dir="./original_data"
)
```

The 24 source task folders used (each with 1000 demos):
- 6 **PnP** tasks: Bottle→Cabinet, Can→Drawer, Cup→Drawer, Milk→Microwave, Potato→Microwave, Wine→Cabinet
- 18 **PosttrainPnPNovel** tasks: Various object transfers between surfaces (Cuttingboard→{Basket,Cardboardbox,Pan,Pot,Tieredbasket}, Placemat→{Basket,Bowl,Plate,Tieredshelf}, Plate→{Bowl,Cardboardbox,Pan,Plate}, Tray→{Cardboardbox,Plate,Pot,Tieredbasket,Tieredshelf})

## Processing

Each merged dataset was created by:

1. **Subsampling**: Randomly sampling N episodes (with seed=42) from each of the 24 task folders
2. **Merging**: Combining all sampled episodes into a single dataset with:
- Re-indexed episode IDs (0 to total-1)
- Globally re-indexed task IDs across all source folders
- Updated parquet files with corrected `episode_index`, `task_index`, and `index` columns
- Consolidated `meta/` files (`info.json`, `episodes.jsonl`, `tasks.jsonl`, `modality.json`)

### Note on task counts

The `100x24` variant has 185 unique tasks instead of 186. This is because some task folders contain multiple task variants (e.g., the `PosttrainPnPNovelFromTrayToTieredbasket` folder has 10 different object variants like eggplant, tomato, etc.). The "eggplant" variant only appears in 58 out of 1000 source episodes, so when randomly sampling only 100 episodes, none of the eggplant episodes were selected. This is expected behavior with random subsampling.

## Dataset Structure (LeRobot v2.0)

```
<variant>/
├── data/
│ └── chunk-000/
│ ├── episode_000000.parquet
│ ├── episode_000001.parquet
│ └── ...
├── videos/
│ └── chunk-000/
│ └── observation.images.ego_view/
│ ├── episode_000000.mp4
│ ├── episode_000001.mp4
│ └── ...
└── meta/
├── info.json
├── episodes.jsonl
├── tasks.jsonl
├── modality.json
├── stats.json
└── relative_stats.json
```

### Features

| Feature | Type | Shape |
|---------|------|-------|
| `observation.images.ego_view` | video (h264) | [256, 256, 3] @ 20fps |
| `observation.state` | float64 | [44] |
| `action` | float64 | [44] |
| `timestamp` | float64 | [1] |
| `next.reward` | float64 | [1] |
| `next.done` | bool | [1] |
| `task_index` | int64 | [1] |
| `episode_index` | int64 | [1] |
| `annotation.human.coarse_action` | int64 | [1] |

Robot: **GR1ArmsAndWaistFourierHands** (44-DOF: arms, hands, legs, neck, waist)

## Usage

```python
# Download a specific variant
from huggingface_hub import snapshot_download

snapshot_download(
repo_id="Joocjun/GR1-Tabletop-Merged-LeRobot",
repo_type="dataset",
allow_patterns="100x24/**", # or "300x24/**" or "1000x24/**"
local_dir="./gr1_merged"
)
```

## License

Apache 2.0 (following the original NVIDIA dataset license)

## Modification: Next-State Hand Actions

Hand action channels (left_hand indices 7-12, right_hand indices 29-34) have been replaced with next-state values:
- `action[t][7:13] = observation.state[t+1][7:13]`
- `action[t][29:35] = observation.state[t+1][29:35]`
- Last timestep repeats current state.
- `stats.json` removed so training auto-regenerates normalization stats.

许可证：Apache-2.0
任务类别：
- 机器人学
标签：
- LeRobot
- GR00T
- GR1
- robocasa
- 仿真
- 操控
样本量范围：10K<n<100K

# GR1桌面合并式LeRobot数据集

本数据集为源自NVIDIA PhysicalAI-Robotics-GR00T-X-Embodiment-Sim数据集集合的GR1桌面操控数据集的合并与下采样版本，采用**LeRobot v2.0**格式。

## 数据集变体

| 数据集变体 | 演示数/任务 | 任务详情 | 总回合数 | 总帧数 | 近似大小 |
|---------|-----------|-------|----------------|--------------|-------------|
| `1000x24/` | 1000 | 24个文件夹，186个独特任务 | 24,000 | 6,020,058 | ~40 GB |
| `300x24/` | 300 | 24个文件夹，186个独特任务 | 7,200 | 1,803,236 | ~12 GB |
| `100x24/` | 100 | 24个文件夹，185个独特任务 | 2,400 | 602,846 | ~4 GB |

## 源数据集

原始单任务数据集下载自：**[nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim](https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim)**

要下载原始数据集：
bash
# 安装huggingface_hub
pip install huggingface_hub

# 下载特定任务数据集
from huggingface_hub import snapshot_download
snapshot_download(
repo_id="nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim",
repo_type="dataset",
allow_patterns="gr1_unified.PnP*", # 或特定任务模式
local_dir="./original_data"
)

所用的24个源任务文件夹（每个包含1000个演示）：
- 6个**抓取放置（Pick-and-Place，简称PnP）**任务：瓶子→储物柜、易拉罐→抽屉、杯子→抽屉、牛奶盒→微波炉、土豆→微波炉、红酒→储物柜
- 18个**训练后新颖PnP（PosttrainPnPNovel）**任务：各类物体在不同台面间的转移任务（切菜板→{篮筐、纸箱、平底锅、锅具、分层篮}、餐垫→{篮筐、碗、盘子、分层架}、盘子→{碗、纸箱、平底锅、盘子}、托盘→{纸箱、盘子、锅具、分层篮、分层架}）

## 处理流程

各合并数据集的构建流程如下：

1. **下采样**：从24个任务文件夹中随机采样N个回合（随机种子设为42）
2. **合并**：将所有采样得到的回合整合为单个数据集，包含：
- 重新索引的回合ID（范围为0至总回合数-1）
- 跨所有源文件夹的全局统一任务索引
- 更新后的parquet文件，修正了`episode_index`、`task_index`和`index`字段
- 整合后的`meta/`目录文件（`info.json`、`episodes.jsonl`、`tasks.jsonl`、`modality.json`）

### 任务数量说明

`100x24`变体的独特任务数为185而非186。这是因为部分任务文件夹包含多种任务变体（例如`PosttrainPnPNovelFromTrayToTieredbasket`文件夹包含茄子、番茄等10种不同的物体变体）。其中“茄子”变体仅在1000个源回合中出现58次，因此当仅随机采样100个回合时，未选中任何茄子变体的回合。这属于随机下采样的正常现象。

## LeRobot v2.0格式的数据集结构

<变体>/
├── data/
│ └── chunk-000/
│ ├── episode_000000.parquet
│ ├── episode_000001.parquet
│ └── ...
├── videos/
│ └── chunk-000/
│ └── observation.images.ego_view/
│ ├── episode_000000.mp4
│ ├── episode_000001.mp4
│ └── ...
└── meta/
├── info.json
├── episodes.jsonl
├── tasks.jsonl
├── modality.json
├── stats.json
└── relative_stats.json

### 数据特征

| 特征 | 类型 | 形状 |
|---------|------|-------|
| `observation.images.ego_view` | 视频（h264编码） | [256, 256, 3] @ 20fps |
| `observation.state` | float64 | [44] |
| `action` | float64 | [44] |
| `timestamp` | float64 | [1] |
| `next.reward` | float64 | [1] |
| `next.done` | 布尔值 | [1] |
| `task_index` | int64 | [1] |
| `episode_index` | int64 | [1] |
| `annotation.human.coarse_action` | int64 | [1] |

机器人构型：**GR1ArmsAndWaistFourierHands**（44个自由度：双臂、双手、双腿、颈部、腰部）

## 使用方法

python
# 下载指定变体
from huggingface_hub import snapshot_download

snapshot_download(
repo_id="Joocjun/GR1-Tabletop-Merged-LeRobot",
repo_type="dataset",
allow_patterns="100x24/**", # 或 "300x24/**" 或 "1000x24/**"
local_dir="./gr1_merged"
)

## 许可证

Apache 2.0（遵循原始NVIDIA数据集许可证）

## 修改内容：下一状态手部动作

手部动作通道（左手索引7-12、右手索引29-34）已替换为下一状态值：
- `action[t][7:13] = observation.state[t+1][7:13]`
- `action[t][29:35] = observation.state[t+1][29:35]`
- 最后一个时间步重复当前状态
- 已移除`stats.json`，训练过程将自动重新生成归一化统计量。