Plan retrieval

This project has been developed with Python 3.12

Setup instructions

Create and activate a virtual environment:

python3 -m venv .venv
source .venv/bin/activate

Install required packages:

python3 -m pip install -r requirements.txt

Run the evaluation script

Download Wordnet for semantic similarity:

python3
> import nltk
> nltk.download('wordnet')

Modify path and configuration:

1. path to LLM-Planning-PlanBench repository in scripts/plan-recovery/check_plan_validity.py
2. name of model used for generation
3. name of results file

Run the evaluation script to run the full pipeline on all model/task combinations:

cd scripts/plan_recovery
./run_experiments.sh

Or run the evaluation script directly for a specific model/task combination:`

cd scripts/plan_recovery
python3 run_evaluation.py --model <model_name> --task <task_name>