yxgao/egodex-wds
收藏Hugging Face2026-03-24 更新2026-03-29 收录
下载链接：
https://hf-mirror.com/datasets/yxgao/egodex-wds
下载链接
链接失效反馈官方服务：
资源简介：
---
task_categories:
- robotics
language:
- en
tags:
- egocentric
- robotics
pretty_name: EgoDex
size_categories:
- 100K<n<1M
---
```python
from datasets import load_dataset, Features, Value, Sequence, Video, Array2D, Array4D
EGODEX_DOF = 69
features = Features({
"__key__": Value("string"),
"__url__": Value("string"),
"confidences.npy": Array2D(shape=(None, EGODEX_DOF), dtype="float32"),
"intrinsic.npy": Array2D(shape=(3, 3), dtype="float32"),
"json": {
"episode_id": Value("int64"),
"has_confidences": Value("bool"),
"joint_names": Sequence(Value("string")),
"llm_description": Value("string"),
"llm_description2": Value("string"),
"num_frames": Value("int64"),
"num_joints": Value("int64"),
"original_resolution": Sequence(Value("int64")),
"target_resolution": Sequence(Value("int64")),
"task": Value("string"),
"which_llm_description": Value("int64"),
},
"mp4": Video(),
"transforms.npy": Array4D(
shape=(None, EGODEX_DOF, 4, 4),
dtype="float32"
),
"txt": Value("string"),
})
dataset = load_dataset("yxgao/egodex-wds", streaming=True, features=features)
```
### 数据集元信息
task_categories:
- 机器人学
language:
- 英语
tags:
- 自我中心视角（egocentric）
- 机器人学
pretty_name: EgoDex
size_categories:
- 100K<n<100万
python
from datasets import load_dataset, Features, Value, Sequence, Video, Array2D, Array4D
# EgoDex数据集的自由度（Degrees of Freedom，DOF）固定为69
EGODEX_DOF = 69
# 定义数据集的特征结构
features = Features({
"__key__": Value("string"),
"__url__": Value("string"),
"confidences.npy": Array2D(shape=(None, EGODEX_DOF), dtype="float32"),
"intrinsic.npy": Array2D(shape=(3, 3), dtype="float32"),
"json": {
"episode_id": Value("int64"),
"has_confidences": Value("bool"),
"joint_names": Sequence(Value("string")),
"llm_description": Value("string"),
"llm_description2": Value("string"),
"num_frames": Value("int64"),
"num_joints": Value("int64"),
"original_resolution": Sequence(Value("int64")),
"target_resolution": Sequence(Value("int64")),
"task": Value("string"),
"which_llm_description": Value("int64"),
},
"mp4": Video(),
"transforms.npy": Array4D(
shape=(None, EGODEX_DOF, 4, 4),
dtype="float32"
),
"txt": Value("string"),
})
# 以流式模式加载yxgao/egodex-wds数据集，并使用上述定义的特征结构
dataset = load_dataset("yxgao/egodex-wds", streaming=True, features=features)
提供机构：
yxgao
