search

OpenDriveLab / UniAD

[CVPR 2023 Best Paper Award] Planning-oriented Autonomous Driving
Apache License 2.0
3.57k stars 406 forks source link

What does "num_groups" mean in the shape (num_groups, num_anchor, 12, 2) of motion_anchor_infos_mode6.pkl? #202

Open anapple-hub opened 2 months ago

anapple-hub commented 2 months ago 
      anchor_infos = pickle.load(open(anchor_info_path, 'rb'))
      self.kmeans_anchors = torch.stack(
          [torch.from_numpy(a) for a in anchor_infos["anchors_all"]])  # Nc, Pc, steps, 2

What does "num_groups" mean in the shape (num_groups, num_anchor, 12, 2) of motion_anchor_infos_mode6.pkl?