train_state.json

{"best_reward": 0.9137129187583923, "last_eval_reward": 0.9137129187583923, "episode": 7510, "saved_at": "2026-01-02T02:34:47.295728"}