Repository
Models / Research
KERV · zhengzihaoPKU
Code of paper "KERV: Kinematic-Rectified Speculative Decoding for Embodied VLA Models" accepted by DAC 2026.
Overview
Code of paper "KERV: Kinematic-Rectified Speculative Decoding for Embodied VLA Models" accepted by DAC 2026.
Review observation/action interfaces, checkpoints, robot embodiments and adaptation requirements. Repository code licensing does not automatically license weights or training data.
Project information
- Maintainer
- zhengzihaoPKU
- Primary language
- Python
- Repository code license
- MIT
- Model-weight license
- Not recorded
- Dataset license
- Not recorded
- Hardware design license
- Not recorded
- Maintenance
- Not archived in the last recorded metadata
- Last observed repository push
- 2026-08-20
Code licenses do not automatically cover model weights, datasets or hardware designs. Unrecorded license details require confirmation with the publisher.
Documentation sections
- News
- Overview
- Method
- KF-Based Compensation Mechanism
- Kinematic-Based Threshold Adjustment
- On-Device Runtime Optimization
Browse this resource’s categories
Explore related collections
Vision-language-action models ↗More in Models
Repository
