pickle_model = load_from_hub(repo_id="ThomasSimonini/q-Taxi-v3", filename="q-learning.pkl")

# Don't forget to check if you need to add additional attributes (is_slippery etc)
env = gym.make(model["env_id"])

evaluate_agent(env, model["max_steps"], model["n_eval_episodes"], model["qtable"], model["eval_seed"])