README.md

metadata

tags:
  - FrozenLake-v1-8x8
  - q-learning
  - reinforcement-learning
  - custom-implementation
model-index:
  - name: q-FrozenLake-v1-8x8-Slippery
    results:
      - task:
          type: reinforcement-learning
          name: reinforcement-learning
        dataset:
          name: FrozenLake-v1-8x8
          type: FrozenLake-v1-8x8
        metrics:
          - type: mean_reward
            value: 0.37 +/- 0.48
            name: mean_reward
            verified: false

Q-Learning Agent playing FrozenLake-v1

This is a trained model of a Q-Learning agent playing FrozenLake-v1 .

Usage

model = load_from_hub(repo_id="jmsalvi/q-FrozenLake-v1-8x8-Slippery", filename="q-learning.pkl")

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

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