README.md

metadata

tags:
  - LunarLander-v2
  - ppo
  - deep-reinforcement-learning
  - reinforcement-learning
  - custom-implementation
  - deep-rl-course
model-index:
  - name: PPO
    results:
      - task:
          type: reinforcement-learning
          name: reinforcement-learning
        dataset:
          name: LunarLander-v2
          type: LunarLander-v2
        metrics:
          - type: mean_reward
            value: 277.79 +/- 24.31
            name: mean_reward
            verified: false

PPO Agent Playing LunarLander-v2

This is a trained model of a PPO agent playing LunarLander-v2.

Hyperparameters

{'exp_name': 'main'
'gym_id': 'LunarLander-v2'
'lr': 0.004234982547109355
'seed': 1
'total_timesteps': 5000000
'torch_deterministic': True
'cuda': True
'num_envs': 32
'num_steps': 256
'anneal_lr': True
'gae': True
'gamma': 0.99
'gae_lambda': 0.8
'num_minibatches': 128
'update_epochs': 20
'norm_adv': False
'clip_coef': 0.1
'vl_clip': True
'ent_coef': 0.00023233247122755093
'vf_coef': 0.6401567434560633
'max_grad_norm': 0.3
'target_kl': 0.012736232712398961
'batch_size': 8192
'minibatch_size': 64
'repo_id': 'Statos6/ppo-cleanRL-LunarLander-v2'
'env_id': 'LunarLander-v2'}