TrajTok: Technical Report for 2025 Waymo Open Sim Agents Challenge
About
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.
Zhiyuan Zhang, Xiaosong Jia, Guanyu Chen, Qifeng Li, Junchi Yan• 2025
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Traffic Simulation | Waymo Open Motion Dataset (WOMD) v2025 (private test) | RMM Score78.61 | 14 |
Showing 1 of 1 rows