eval_test_LAVLA_async_inference_v2_2026_06_11
收藏Hugging Face2026-06-11 更新2026-06-11 收录
下载链接:
https://huggingface.co/datasets/Alkatt/eval_test_LAVLA_async_inference_v2_2026_06_11
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是一个机器人控制数据集,使用LeRobot工具创建,专为机器人学习和评估设计。数据集包含一个完整的episode,总计560帧,数据以parquet格式存储。主要特征包括:动作数据(action),表示6个关节位置(肩部平移、肩部提升、肘部弯曲、手腕弯曲、手腕旋转和夹爪位置);状态观测(observation.state),同样为6个关节位置;以及来自三个摄像头(camera1、camera2、camera3)的图像观测,每个视频分辨率为480x640,帧率为30fps,编码为av1格式。此外,数据集还包含时间戳、帧索引、episode索引、任务索引等元数据。该数据集适用于机器人模仿学习、强化学习或视觉控制任务,数据规模为100MB的数据文件和200MB的视频文件,所有数据按Apache 2.0许可证授权。
This robotic control dataset was constructed using the LeRobot toolkit, and is specifically designed for robot learning and evaluation. It contains one complete episode with a total of 560 frames, with all data stored in Parquet format. The main features of the dataset include: action data (`action`), which represents the positions of 6 joints including shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist rotation and gripper position; state observation (`observation.state`), which also consists of 6 joint position values; and image observations from three cameras (camera1, camera2, camera3), where each video has a resolution of 480×640, a frame rate of 30 fps, and is encoded in AV1 format. In addition, the dataset includes metadata such as timestamps, frame indices, episode indices and task indices. This dataset is applicable to robotic imitation learning, reinforcement learning or visual control tasks. The total dataset comprises a 100 MB data file and a 200 MB video file, and all data is licensed under the Apache 2.0 license.
提供机构:
Alkatt创建时间:
2026-06-11
原始信息汇总
数据集概述
- 数据集名称: eval_test_LAVLA_async_inference_v2_2026_06_11
- 创建工具: 使用 LeRobot 框架创建
- 许可证: Apache-2.0
- 任务类型: 机器人学 (robotics)
- 标签: LeRobot
数据集规模
- 总片段数 (episodes): 1
- 总帧数 (frames): 560
- 总任务数 (tasks): 1
- 数据文件大小: 100 MB
- 视频文件大小: 200 MB
- 帧率 (FPS): 30
- 块大小 (chunks size): 1000
数据集结构
数据类型
- 采用 Parquet 文件格式存储数据,视频以 AV1 编码的 MP4 文件存储。
- 数据路径模板:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频路径模板:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4
特征字段
| 特征名 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
action |
float32 | (6,) | 6维动作指令(机器人关节位置) |
observation.state |
float32 | (6,) | 6维观测状态(机器人关节位置) |
observation.images.camera1 |
video | (480, 640, 3) | 相机1视频(480x640,RGB) |
observation.images.camera2 |
video | (480, 640, 3) | 相机2视频(480x640,RGB) |
observation.images.camera3 |
video | (480, 640, 3) | 相机3视频(480x640,RGB) |
timestamp |
float32 | (1,) | 时间戳 |
frame_index |
int64 | (1,) | 帧索引 |
episode_index |
int64 | (1,) | 片段索引 |
index |
int64 | (1,) | 索引 |
task_index |
int64 | (1,) | 任务索引 |
动作/状态命名空间
shoulder_pan.posshoulder_lift.poselbow_flex.poswrist_flex.poswrist_roll.posgripper.pos
视频编码信息
- 宽度: 640 像素
- 高度: 480 像素
- 编码器: AV1
- 像素格式: yuv420p
- 帧率: 30 FPS
- 通道数: 3 (RGB)
- 深度图: 否
数据集划分
- 训练集:
0:1(全部数据)
机器人信息
- 机器人类型:
so_follower(跟随机器人)



