generalization ability for new forece direction
TingtingLiao opened this issue · comments
Hi, thanks for your excellent work. I have a question regarding force direction. When I applied a new direction not included in the training parameters, the results look wired. Could you provide some guidance on this issue?
inference_demo_carnations_sv_gres64_substep768_model_0_force_2_mag_2.0_point_0_movingcamera_camid_0_inference_demo_carnations_sv_gres64_substep768_model_0_force_2_mag_2.0_point_0.mp4
Hi, TingtingLiao,
The videos seems to be a result of some bug. Which command are you using?
Also, can you try the same code with some other forces, and see if their results is also wired?