mbodiai/ABB_PandG_20250402_112504

这个数据集包含了机器人操作的相关信息，其中包括观测到的图像(image)、指令(instruction)和提示(prompt)；动作(action)包括姿态(pose)、抓握(grasp)和夹爪力(gripper_force)；状态(state)包括前后摄像头图像(camF, camT)、深度信息(depths)、世界中的物体(world_objects)和夹爪姿态(gripper_pose)；奖励(reward)是浮点数类型；元数据(metadata)包括剧集索引(episode_idx)和步索引(step_idx)。训练集(train)共有12个示例，大小为126820972.0字节。

This dataset contains information related to robotic manipulation, including observed images (image), instructions (instruction), and prompts (prompt); actions (action) include pose, grasp, and gripper force; state (state) includes front and rear camera images (camF, camT), depth information (depths), objects in the world (world_objects), and gripper pose; reward is a floating-point number; metadata includes episode index (episode_idx) and step index (step_idx). The training set (train) has 12 examples and is 126820972.0 bytes in size.