angle_peg_stereo_05_27_tip_ref
收藏资源简介:
该数据集使用LeRobot框架创建,是一个机器人学领域的数据集,适用于机器人模仿学习或强化学习任务。数据集以Parquet文件格式存储,包含多模态观测数据和对应的机器人动作指令。观测数据包括来自三个摄像头(cam0, cam1, cam2)的图像,每个图像为3通道、240x320分辨率;以及一个3维的机器人状态向量(observation.state)。动作数据(action)是一个10维的浮点数向量。此外,数据集中还包含时间戳(timestamp)、帧索引(frame_index)、回合索引(episode_index)、全局索引(index)和任务索引(task_index)等元数据。数据集规模为72个完整的训练回合(episodes),共计44185帧数据,原始数据文件大小约为100MB,对应的视频文件大小约为200MB。数据采集帧率为15帧/秒,使用的机器人平台为Franka。数据集已按照训练集(所有72个回合)进行了划分。
This dataset is developed using the LeRobot framework, belonging to the robotics domain and suitable for robot imitation learning and reinforcement learning tasks. Stored in Parquet file format, it contains multimodal observation data and corresponding robot action instructions. The observation data includes images from three cameras (cam0, cam1, cam2), each with 3 channels and a resolution of 240×320, as well as a 3-dimensional robot state vector (observation.state). The action data (action) is a 10-dimensional floating-point vector. In addition, the dataset also includes metadata such as timestamp, frame_index, episode_index, global index, and task_index. The dataset consists of 72 complete training episodes, totaling 44,185 frames of data. The raw data file is approximately 100 MB in size, while the corresponding video files total around 200 MB. The data was collected at a frame rate of 15 frames per second, using the Franka robotic platform. The dataset has been partitioned exclusively for training using all 72 episodes.
- 许可证: Apache-2.0
- 任务类别: 机器人学 (Robotics)
- 标签: LeRobot
数据集描述
该数据集使用 LeRobot 创建,包含机器人操作相关的数据。
数据集结构
- 代码库版本: v3.0
- 帧率 (FPS): 15
- 机器人类型: franka
- 总片段数: 72
- 总帧数: 44185
- 总任务数: 1
- 数据块大小: 1000
- 数据文件大小: 100 MB
- 视频文件大小: 200 MB
- 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 - 拆分 (Splits):
- 训练集 (train): 0 到 72 片段
特征 (Features)
| 特征名 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
observation.image.cam0 |
图像 (image) | [3, 240, 320] | 摄像头0图像 (通道, 高度, 宽度) |
observation.image.cam1 |
图像 (image) | [3, 240, 320] | 摄像头1图像 (通道, 高度, 宽度) |
observation.image.cam2 |
图像 (image) | [3, 240, 320] | 摄像头2图像 (通道, 高度, 宽度) |
observation.state |
float32 | [3] | 机器人状态 (维度: D_state) |
action |
float32 | [10] | 动作 (维度: D_act) |
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 |




