mbodiai/bop-shapenet-mega-pose-999

该数据集包含多个特征，如图像、episode_idx、step_idx、timestamp、observation、action、image_keys和from_lists。observation特征进一步包含camera和image子特征，camera特征又包含depth_scale、distortion、extrinsic和intrinsic子特征。action特征包含objects子特征，objects子特征进一步包含bbox_2d、bbox_3d、mask、name、pixel_coords、pose和volume子特征。数据集分为train分割，包含999个样本，总大小为981406590字节。

The dataset contains multiple features such as image, episode_idx, step_idx, timestamp, observation, action, image_keys, and from_lists. The observation feature further includes camera and image sub-features, with the camera feature containing depth_scale, distortion, extrinsic, and intrinsic sub-features. The action feature includes objects sub-feature, which further contains bbox_2d, bbox_3d, mask, name, pixel_coords, pose, and volume sub-features. The dataset is divided into a train split, containing 999 samples, with a total size of 981406590 bytes.