| # Examples |
|
|
| Self-contained scripts that exercise the dataset end-to-end. They all |
| work against the downloaded CSV files only — no extra dependencies |
| beyond `numpy` and the Python standard library. |
|
|
| | Script | What it does | |
| |---|---| |
| | `baseline_constant_velocity.py` | Loads one split, predicts future positions by extrapolating the last-step velocity, prints ADE and FDE. The simplest fair baseline. | |
|
|
| For env-aware or social-aware models, see the trained checkpoints in |
| `checkpoints/` and the model definitions referenced in the v1 paper |
| (repository code at https://github.com/mark000071/envship_v2_datasets). |
|
|
