{ "timesteps": 3000000, "learning_rate": 3e-4, "batch_size": 256, "n_envs": 16, "n_steps": 2048, "gamma": 0.99, "ent_coef": 0.1, "n_epochs": 10, "gae_lambda": 0.95, "clip_range": 0.2, "vf_coef": 0.5, "max_grad_norm": 0.5, "normalize_advantage": true }