README.md

metadata

tags:
  - HealthGatheringSupreme-v1
  - ppo
  - deep-reinforcement-learning
  - reinforcement-learning
  - custom-implementation
  - deep-rl-course
  - sample-factory
model-index:
  - name: PPO
    results:
      - task:
          type: reinforcement-learning
          name: reinforcement-learning
        dataset:
          name: doom_health_gathering_supreme
          type: doom_health_gathering_supreme
        metrics:
          - type: mean_reward
            value: 18.30 +/- 8.82
            name: mean_reward
            verified: false

PPO Agent Playing HealthGatheringSupreme-v1

This is a trained model of a PPO agent playing HealthGatheringSupreme-v1 using a custom CleanRL PPO implementation (not sample factory).

Hyperparameters

{'env_id': 'HealthGatheringSupreme-v1'
'learning_rate': 0.0001
'learning_rate_min': 1e-06
'gamma': 0.99
'gae_lambda': 0.95
'clip_coef': 0.2
'total_timesteps': 10000000
'recurrence': 32
'ent_coef': 0.001
'vf_coef': 0.5
'max_grad_norm': 0.5
'num_minibatches': 4
'update_epochs': 1
'frame_skip': 4}