eval_LAVLA_Async_VI_100M_40k_b20_whiteblackcube_v2_2026_05_27_white_latency_bench_v5
收藏资源简介:
该数据集由LeRobot创建,是一个专用于机器人学研究的机器人操作数据集。它包含2个episodes、933个帧和1个任务,数据以parquet格式存储。数据集的特征涵盖6维浮点型动作向量,用于控制机器人的肩部、肘部、腕部和夹爪位置;6维浮点型观察状态向量,表示与动作相同的关节位置;以及三个摄像头(camera1、camera2、camera3)的视频观察,每个视频分辨率为480x640,3通道,帧率为30fps。此外,数据集还包含时间戳、帧索引、episode索引、索引和任务索引等元数据。数据集采用Apache 2.0许可证,适用于机器人控制、模仿学习或强化学习等任务。
This dataset is created by LeRobot and is designed for robotics research, specifically as a robot manipulation dataset. It contains 2 episodes, 933 frames, and 1 task, with data stored in the parquet format. The features include a 6-dimensional floating-point action vector for controlling the positions of the shoulder, elbow, wrist, and gripper; a 6-dimensional floating-point observation state vector representing the same joint positions as the actions; video observations from three cameras (camera1, camera2, camera3), each with a resolution of 480x640, 3 channels, and a frame rate of 30fps; and metadata such as timestamps, frame indices, episode indices, indices, and task indices. The dataset is licensed under Apache 2.0 and is suitable for tasks like robot control, imitation learning, or reinforcement learning.
数据集概述
- 名称:
Alkatt/eval_LAVLA_Async_VI_100M_40k_b20_whiteblackcube_v2_2026_05_27_white_latency_bench_v5 - 许可证: Apache-2.0
- 任务类别: 机器人学 (robotics)
- 标签: LeRobot
- 创建工具: 使用 LeRobot 创建
数据集规模与结构
- 总片段数: 2
- 总帧数: 933
- 总任务数: 1
- 帧率 (FPS): 30
- 数据文件大小: 100 MB
- 视频文件大小: 200 MB
- 分块大小: 1000
- 数据路径:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频路径:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4 - 机器人类型: so_follower
- 代码库版本: v3.0
- 数据集划分: 仅包含训练集 (train),对应全部片段 (0:2)
特征详情
-
动作 (action)
- 数据类型: float32
- 维度: 6
- 名称:
shoulder_pan.pos,shoulder_lift.pos,elbow_flex.pos,wrist_flex.pos,wrist_roll.pos,gripper.pos
-
观测状态 (observation.state)
- 数据类型: float32
- 维度: 6
- 名称: 与动作相同
-
观测图像 (observation.images)
- 包含三个摄像头:
camera1,camera2,camera3 - 每个摄像头的图像尺寸: 480×640 像素,3 通道
- 视频编码: av1
- 像素格式: yuv420p
- 帧率: 30 FPS
- 非深度图
- 包含三个摄像头:
-
时间戳 (timestamp)
- 数据类型: float32
- 维度: 1
-
帧索引 (frame_index)
- 数据类型: int64
- 维度: 1
-
片段索引 (episode_index)
- 数据类型: int64
- 维度: 1
-
索引 (index)
- 数据类型: int64
- 维度: 1
-
任务索引 (task_index)
- 数据类型: int64
- 维度: 1



