ppo_v82/policy_update10.state.json

{
  "checkpoint_schema_version": 2,
  "hidden_reset_policy": "episode_done",
  "update": 11,
  "stage": 5,
  "ent_scale": 1.0,
  "lr_now": 0.00025,
  "total_env_steps": 22528,
  "recent_wins": [],
  "best_eval_win": null,
  "best_eval_score": null,
  "v81_curriculum": true,
  "curriculum_schedule": "v8.2",
  "stage_env_targets": [
    24,
    24,
    24,
    24,
    24,
    16,
    12,
    10,
    12,
    10,
    8,
    8,
    8,
    8
  ],
  "envs_per_shard": 16,
  "requested_env_target": null
}