About repr_abs_only during the test on PROX/Egobody datasets

[JiwanHur]

In the paper, it is mentioned "excluding first derivatives to avoid global drifting caused by inaccurate velocities." (i.e. repr_abs_only=True)

With this setting, it is impossible to reconstruct the whole root trajectory vector because the relative trajectory information is not available when applying the trajectory mask to the prox or egobody datasets. In the 'test_prox_egobody.py', it seems that the trajectory mask is not applied and the noisy relative trajectory information is directly applied.

1. Do the initialized vectors from the prox or egobody datasets not have the trajectory mask?
2. If so, doesn't it harm the performance of the reconstruction when directly using the noisy relative trajectory information?