Alkatt/eval_LAVLA_Async_VI_100M_40k_b20_whiteblackcube_v2_2026_05_27_white_latency_bench_v12
收藏资源简介:
该数据集是使用LeRobot创建的机器人数据集,专为机器人学习任务设计。数据集包含一个名为so_follower的机器人类型,总共有10个episodes和8350帧数据。数据以parquet文件格式存储,分为多个块(chunks),每个块大小为1000帧。数据集包括动作数据(如肩部、肘部、腕部和抓取器的位置)、状态观测数据(与动作相同),以及来自三个摄像头的图像观测数据(每个摄像头分辨率为480x640,帧率30fps,视频编码为av1)。此外,还包含时间戳、帧索引、episode索引、任务索引等元数据。数据集总数据文件大小为100MB,视频文件大小为200MB,适用于训练和评估机器人控制模型。
This dataset was created using LeRobot and is designed for robotics learning tasks. It features a robot type called so_follower, with a total of 10 episodes and 8350 frames. The data is stored in parquet format, organized into chunks of 1000 frames each. The dataset includes action data (e.g., positions of shoulder, elbow, wrist, and gripper), state observation data (same as actions), and image observations from three cameras (each with a resolution of 480x640, 30fps, and av1 video codec). Additionally, it contains metadata such as timestamps, frame indices, episode indices, and task indices. The total data file size is 100MB, and the video file size is 200MB, making it suitable for training and evaluating robot control models.




