Tsagkas/dataset_deltas_lvfm_v10

该数据集使用LeRobot创建，是一个机器人领域的数据集，专门针对so101_follower机器人类型。数据集包含48个训练片段，总计6685帧，涉及2个任务，数据以parquet文件（总大小100MB）和视频文件（总大小200MB）格式存储，帧率为10fps。数据集结构包括动作特征（如末端执行器的位置、旋转和夹爪状态）、观测状态特征（包括手腕和前置摄像头图像、VFM/VLM特征、技能特征、对象特征和全文本特征），以及时间戳、帧索引、片段索引和任务索引等元数据。所有数据仅用于训练，无其他分割。

This dataset was created using LeRobot and belongs to the robotics domain, specifically for the so101_follower robot type. It contains 48 training episodes with a total of 6685 frames, covering 2 tasks. The data is stored in parquet files (total size 100MB) and video files (total size 200MB) with a frame rate of 10fps. The dataset structure includes action features (such as end-effector position, rotation, and gripper state), observation state features (including wrist and front camera images, VFM/VLM features, skill features, object features, and full-text features), as well as metadata like timestamp, frame index, episode index, and task index. All data is for training purposes only, with no other splits.