the dimension of pose for the action modality appears to be 24 not 27
#2
by
zhaonat - opened
seems to be the case for most of the data I've checked.
Hi, thanks for the observation. The discrepancy is because the Action modality only has 24 dimensions, as we do not record the head's position signal (since it remains stationary during actions). However, the State modality includes the head data, totaling 27 dimensions.
epoch2000 changed discussion status to
closed