LaWAM Pretrained Policy for LeRobot
This repository contains the LaWAM pretrained policy converted to the native
LeRobot checkpoint format. Model parameters are stored in model.safetensors;
loading this repository does not require the original PyTorch .pt checkpoint.
- Paper: LaWAM: Latent World Action Models for Efficient Dynamics-Aware Robot Policies
- Project page: rlinf.github.io/LaWAM
- Source code: RLinf/LaWAM
- Original checkpoint: jialei02/lawam_pretrain
- LeRobot integration: huggingface/lerobot#3999
Intended Use
This is the pretrained LaWAM policy intended as an initialization checkpoint for downstream fine-tuning. The serialized default feature schema and processor statistics use the 7-dimensional Franka setup from the original release. Other robot embodiments require the downstream LeRobot dataset feature schema and normalization statistics to be supplied during fine-tuning.
The policy uses Qwen/Qwen3-VL-2B-Instruct as its portable base VLM identifier.
The Qwen model is downloaded automatically unless base_vlm_path is overridden
with a local copy.
Loading
LaWAM support requires the LeRobot integration from PR #3999 or a later LeRobot release containing that integration.
from lerobot.policies.lawam import LaWAMPolicy
policy = LaWAMPolicy.from_pretrained("jialei02/lawam-pretrain-lerobot")
The checkpoint includes the LeRobot policy config, preprocessor, postprocessor, normalization tensors, and conversion provenance manifest.
Conversion Verification
The source checkpoint, source config, dataset statistics, LAM architecture
config, and Qwen config/model were checked against their published SHA-256
values before conversion. A strict local from_pretrained() reload restored all
1,912 policy state keys while torch.load was disabled.
Citation
@misc{chen2026lawam,
title = {LaWAM: Latent World Action Models for Efficient Dynamics-Aware Robot Policies},
author = {Chen, Jialei and Wang, Kai and Chen, Kang and Chen, Shuaihang and Gao, Feng and Tang, Wenhao and Li, Zhiyuan and Liu, Weilin and Yao, Zhuyu and Li, Boxun and Xu, Yuanbo and Yu, Chao},
journal = {arXiv preprint arXiv:2606.15768},
year = {2026},
archiveprefix = {arXiv},
primaryclass = {cs.RO},
}