{ "hidden_sizes": [ 64, 64 ], "activation": "tanh", "learning_rate": 0.0003, "clip_epsilon": 0.2, "gae_lambda": 0.95, "gamma": 0.99, "entropy_coef": 0.01, "value_coef": 0.5, "max_grad_norm": 0.5, "n_epochs": 10, "batch_size": 64, "rollout_steps": 52, "total_timesteps": 10000, "cost_bps": 10, "cash_buffer": 0.02, "max_position_weight": 0.2, "reward_scale": 100.0, "n_stocks": 15, "seed": 42, "validation_years": 2, "min_sharpe_improvement": 0.0, "patchtst_version": null, "training_years": 10, "n_eval_folds": 3 }