Downloads · 30 days
0
pls-vla-skill-predictor
erl-hub/pls-vla-skill-predictor
pls-vla-skill-predictor is a robotics model from erl-hub. Use it for the robotics task on the model card, and read the license before you ship it in a product. It is set up for pls-vla.
Transformer Skill Predictor SPθ(V(ft), et) → (Ŝt, p̂t). Checkpoint folders use the same names as W&B runs: sp-<dataset (e.g. sp-b1k-task0000).
Access
Public
Updated Aug 5, 2026
Repo size
17.8 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt17.8 MB · 100%
At a glance
- Task
- Robotics
- Library
- pls-vla
- Access
- Public
- Created
- Aug 5, 2026
- Updated
- Aug 5, 2026
- SHA
- dd254e17
Model card
From the Hugging Face model README
PLS-VLA Skill Predictor
Transformer Skill Predictor SP_θ(V(f_t), e_t) → (Ŝ_t, p̂_t).
Checkpoint folders use the same names as W&B runs: sp-<dataset>
(e.g. sp-b1k-task0000).
Checkpoint sp-b1k-task0000
- file:
sp-b1k-task0000/sp.pt - d_vis=768 d_pre=4096 d_skill=4096
- d_model=256 layers=4 heads=8
- residual=False
- trained epochs=20 final train loss=26.258141210804816
- extra:
{"data": "skill_dataset version b1k-task0000 split=train"}
python scripts/run_skill_predictor.py \
--checkpoint <local.pt> # or download sp-b1k-task0000/sp.pt from this repo
Checkpoints in this repo
Folder names match W&B (sp-<dataset>).
sp-b1k-task0000
- Task
- Robotics
- Library
- pls-vla
- Created
- Aug 5, 2026
- Updated
- Aug 5, 2026