Skip to content

AGENTS.md

Latest commit

 

History

History
30 lines (23 loc) · 984 Bytes

AGENTS.md

File metadata and controls

30 lines (23 loc) · 984 Bytes

AGENTS.md

Quick guide for working in this repo.

Project summary

  • distill-lagat: simplified LaGAT (GNN policy + LaCAM search) for MAPF.
  • Python 3.10 + PyTorch (GNN), C++17 (search).

Key paths

  • src/: Python source code.
  • scripts/: CLI entry points for data, training, evaluation.
  • cpp_planners/: C++ planners (LaCAM/LaGAT).
  • assets/: maps, pretrained models, demo artifacts.
  • outputs/: generated datasets and run outputs.
  • tests/: pytest-based checks for features/model.

Typical workflow

# 1) Collect expert trajectories
uv run scripts/collect_expert_trajectories.py num_samples=10 save_animation=True

# 2) Convert to imitation dataset
uv run scripts/convert_to_imitation_dataset.py dataset_dir=/path/to/dataset/

# 3) Train
uv run scripts/train.py dataset_dir=/path/to/imitation_learning_dataset/ num_epochs=10

# 4) Evaluate pretrained model
uv run scripts/eval_model.py model.fpath=assets/pretrained/success_best.jit save_animation=True