---
license: mit
tags:
- robotics
- reinforcement-learning
---

# Play2Perfect checkpoints

State-based policy checkpoints for [Play2Perfect](https://github.com/kushal2000/play2perfect)
(arXiv: 2606.26428).

- `play/model.pth` — Stage-1 "play" pretrained policy.
- `tight_insertion/model.pth` — Stage-2: tight insertion (L-peg, 0.5 mm).
- `beam_assembly_step1/model.pth` — Stage-2: beam assembly step 1.
- `beam_assembly_step2/model.pth` — Stage-2: beam assembly step 2.
- `screwing/model.pth` — Stage-2: screwing (furniture leg).

Download with `python download_checkpoints.py` from the Play2Perfect repo.