Question about new hand urdf

[Keyustar]

While testing the URDF files of the Inspire Hand and Shadow Hand in hand.py, I found that the last two joints of all fingers failed to respond to control via set_drive_target(even though joint.qpos_targetwas updated in the simulation), but their states could be forcibly set using set_qpos.Could you please advise what might be causing this issue?

[Keyustar]

I created a similar visual validation in IsaacGym using the same URDF , but the issue above didn't occur. Is this possibly a bug in Sapien?

[zhaohengyin]

Thanks for the issue. I will look into this. It is likely to be a Sapien simulation problem (this should be only affecting visualization though)

[wczsjtu]

Hello, did you find your training with Inspire URDF effective? My training results are particularly poor.

[Keyustar]

Hello, did you find your training with Inspire URDF effective? My training results are particularly poor.

Sorry, it's been a long time, so I don't remember the training performance of the Inspire Hand clearly. But I'm certain this method can be fully applied to other dexterous hands.

[wczsjtu]

Hello, did you find your training with Inspire URDF effective? My training results are particularly poor.

Sorry, it's been a long time, so I don't remember the training performance of the Inspire Hand clearly. But I'm certain this method can be fully applied to other dexterous hands.

ok,thanks for your reply