Function traj_selection seems has some bugs

Henry1iu / TNT-Trajectory-Prediction

A Unofficial Pytorch Implementation of TNT: Target-driveN Trajectory Prediction

504 stars 95 forks source link

Closed asanasdake closed 2 years ago

asanasdake commented 2 years ago

Hi,

the threshold is halved for each while loop, it will soon fail to filter anything after a few while loops
traj_selected = traj_pred[:, :self.k] should be traj_selected = traj_pred[:, :self.k] .clone()

Henry1iu commented 2 years ago

Hi,

Yes, you are right. Thanks for raising these issues. I will fix these bugs. @

Best Regards, Jianbang