Hi there! I would like to use the SMPL-H data from GRAB that is provided through AMASS, but since I need some of the additional information from the original GRAB dataset, I need to correlate the sequences from GRAB and GRAB-AMASS. I have found that there are some files in GRAB-AMASS without correspondence in GRAB?
In particular, these tasks are present in GRAB-AMASS, but not in the original GRAB dataset:
No corresponding task for banana_eat_1 in s5 found
No corresponding task for phone_lift in s5 found
No corresponding task for cylindersmall_lift in s5 found
No corresponding task for airplane_fly_1 in s5 found
No corresponding task for multi_bottle_whineglass_1 in s3 found
Another note is that the "pick_all" task in GRAB-AMASS seems to correspond to the "lift" task in GRAB?
I would love some clarification on where this extra data is coming from and whether it can safely be ignored, and whether the name change from "lift" to "pick_all" is correct?
Hi there! I would like to use the SMPL-H data from GRAB that is provided through AMASS, but since I need some of the additional information from the original GRAB dataset, I need to correlate the sequences from GRAB and GRAB-AMASS. I have found that there are some files in GRAB-AMASS without correspondence in GRAB?
In particular, these tasks are present in GRAB-AMASS, but not in the original GRAB dataset:
No corresponding task for banana_eat_1 in s5 found No corresponding task for phone_lift in s5 found No corresponding task for cylindersmall_lift in s5 found No corresponding task for airplane_fly_1 in s5 found No corresponding task for multi_bottle_whineglass_1 in s3 found
Another note is that the "pick_all" task in GRAB-AMASS seems to correspond to the "lift" task in GRAB?
I would love some clarification on where this extra data is coming from and whether it can safely be ignored, and whether the name change from "lift" to "pick_all" is correct?