training_config.json

{
  "experiment_name": "ppo_lunarlander_128x128",
  "env_id": "LunarLander-v3",
  "env_kwargs": {},
  "total_timesteps": 5000000,
  "n_envs": 16,
  "seed": 42,
  "actor_layers": [
    128,
    128
  ],
  "critic_layers": [
    128,
    128
  ],
  "learning_rate": 0.0003,
  "n_steps": 1024,
  "batch_size": 64,
  "n_epochs": 4,
  "gamma": 0.999,
  "gae_lambda": 0.98,
  "ent_coef": 0.01,
  "clip_range": 0.2,
  "eval_every_timesteps": 50000,
  "checkpoint_every_timesteps": 100000,
  "n_eval_episodes": 20,
  "device": "cpu"
}