Views
No views yet
GPSStateEstModel — graph-based (GNN + Transformer) state estimation for
variable-vertex cloth meshes. Reconstructs the full cloth mesh state from a
partial point cloud observation via DDPM diffusion, conditioned on each cloth's
own rest state and topology (no global template).fold_unfold_lift_seed_1397.h5)pcdpcd (legacy)pcd rest-position centering convention:
the per-cloth rest (template) positions fed to the model are centered by
subtracting the centroid of the observed point cloud for the same frame.model.config.rest_pos_centering and do this automatically
(checkpoints without the field default to pcd).rest_pos_centering: self
(e.g. fold-unfold-lift-state-est-gps-flow-v2 or
fold-unfold-lift-state-est-gps-sequential).cross_attn_mode: parallel. This checkpoint predates the option (added to the state-estimation model on 2026-05-13 and to the dynamics model on 2026-08-18) and was trained with the original parallel fusion of self- and cross-attention. The key was backfilled into model/config.json on 2026-08-19 so that from_pretrained cannot silently pick another mode — the modes share parameters, so a mismatch loads without error but runs a forward pass the model was never trained with.model/config.json keys; they were added on 2026-08-19 with the values the model was trained with, so from_pretrained no longer depends on the training repo's defaults:rest_pos_centering: pcd — checkpoints predating the field centred the rest positions on the point-cloud centroid.