English

TrajTok: Technical Report for 2025 Waymo Open Sim Agents Challenge

Computation and Language 2025-06-30 v1 Artificial Intelligence

Abstract

In this technical report, we introduce TrajTok, a trajectory tokenizer for discrete next-token-prediction based behavior generation models, which combines data-driven and rule-based methods with better coverage, symmetry and robustness, along with a spatial-aware label smoothing method for cross-entropy loss. We adopt the tokenizer and loss for the SMART model and reach a superior performance with realism score of 0.7852 on the Waymo Open Sim Agents Challenge 2025. We will open-source the code in the future.

Keywords

Cite

@article{arxiv.2506.21618,
  title  = {TrajTok: Technical Report for 2025 Waymo Open Sim Agents Challenge},
  author = {Zhiyuan Zhang and Xiaosong Jia and Guanyu Chen and Qifeng Li and Junchi Yan},
  journal= {arXiv preprint arXiv:2506.21618},
  year   = {2025}
}
R2 v1 2026-07-01T03:35:08.742Z