Project-ML

An implementation of standard ML algorithms applied to taxy v3 of gymnasium Machine Learning Project - Taxi-v3 (Q-learning, DQN, Double DQN)

Setup:

1. crea env (opzionale con conda): conda env create -f environment.yml conda activate taxi_rl oppure: pip install -r requirements.txt

2. struttura project_root come descritto.

Eseguire singoli training: python src/train_q.py --config config.yaml --seed 0 python src/train_dqn.py --config config.yaml --seed 0 python src/train_double_dqn.py --config config.yaml --seed 0

Eseguire tutti i run (multi-seed) e salvare summary: python src/run_experiments.py --config config.yaml

Generare figura comparativa: python src/results_aggregator.py

Valutare un modello salvato: python src/evaluate.py --config config.yaml --path results/dqn/seed_0 --method dqn --episodes 200