Data Quality Issue: Invalid EEF Poses in benchmark1_1_compressed/h5_ur_1rgb

#8
by yangyandan - opened

Hi,

We recently identified a data integrity issue in the benchmark1_1_compressed/h5_ur_1rgb folder regarding invalid End-Effector (EEF) poses.

EEF Pose: The values in puppet/end_effector remain constant throughout the entire trajectory in most episodes.
Joint State: In contrast, the joint position data puppet/joint_position appears valid and varies reasonably.

Could you confirm if this dataset is intended to be used exclusively with joint-space representations? We want to ensure we are not misinterpreting the data format before proceeding with our experiments.
Thanks for your clarification.

Sign up or log in to comment