README.md

metadata

tags:
  - HalfCheetahBulletEnv-v0
  - 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: HalfCheetahBulletEnv-v0
          type: HalfCheetahBulletEnv-v0
        metrics:
          - type: mean_reward
            value: '-1605.41 +/- 93.38'
            name: mean_reward
            verified: false

PPO Agent Playing HalfCheetahBulletEnv-v0

This is a trained model of a PPO agent playing HalfCheetahBulletEnv-v0.

Hyperparameters

{'exp_name': 'parse_args'
'seed': 1
'torch_deterministic': True
'cuda': True
'track': True
'wandb_project_name': 'PPO-RL'
'wandb_entity': None
'capture_video': False
'env_id': 'HalfCheetahBulletEnv-v0'
'total_timesteps': 2000
'learning_rate': 0.0003
'num_envs': 1
'num_steps': 2048
'anneal_lr': True
'gae': True
'gamma': 0.99
'gae_lambda': 0.95
'num_minibatches': 32
'update_epochs': 10
'norm_adv': True
'clip_coef': 0.2
'clip_vloss': True
'ent_coef': 0.0
'vf_coef': 0.5
'max_grad_norm': 0.5
'target_kl': 0.02
'repo_id': 'kismet163/ppo-LunarLander-v2'
'batch_size': 2048
'minibatch_size': 64}