| { | |
| "env": { | |
| "name": "tensoraerospace.envs.b747_vec_torch.ImprovedB747VecEnvTorch", | |
| "params": { | |
| "num_envs": 128, | |
| "dt": 0.1, | |
| "tn": 20.0, | |
| "initial_state": [ | |
| 0.0, | |
| 0.0, | |
| 0.0, | |
| 0.0 | |
| ], | |
| "device": "mps", | |
| "seed": 0, | |
| "auto_reset": true, | |
| "include_reference_in_obs": true, | |
| "step_randomization": { | |
| "signal_type": "step", | |
| "amplitude_deg_range": [ | |
| -5.0, | |
| 5.0 | |
| ], | |
| "min_abs_amplitude_deg": 0.1, | |
| "step_time_sec_range": [ | |
| 1, | |
| 15 | |
| ], | |
| "frequency_hz_range": [ | |
| 0.02, | |
| 0.1 | |
| ], | |
| "p_step": 0.5, | |
| "p_sine": 0.3, | |
| "ref_change_threshold_rad": 0.0017453292519943296, | |
| "min_step_amp_rad": 0.008726646259971648 | |
| }, | |
| "reward_mode": "tracking", | |
| "survival_bonus": 0.0, | |
| "completion_bonus": 0.0, | |
| "early_termination_penalty": 0.0, | |
| "early_termination_penalty_per_step": 0.0 | |
| } | |
| }, | |
| "policy": { | |
| "name": "tensoraerospace.agent.dsac.dsac_flight.DSAC", | |
| "params": { | |
| "gamma": 0.995, | |
| "tau": 0.005, | |
| "alpha": 0.03246142342686653, | |
| "updates_per_step": 4, | |
| "target_update_interval": 1, | |
| "batch_size": 256, | |
| "automatic_entropy_tuning": true, | |
| "target_entropy_scale": 1.0, | |
| "min_alpha": 0.0, | |
| "exploration_noise_std": 0.0, | |
| "reward_clip": 20.0, | |
| "num_quantiles": 8, | |
| "num_quantiles_exp": 8, | |
| "embedding_dim": 64, | |
| "hidden_layers": [ | |
| 64, | |
| 64 | |
| ], | |
| "huber_threshold": 1.0, | |
| "learning_starts": 100000, | |
| "warmup_action_scale": 1.0, | |
| "caps_lambda_smoothness": 0.0, | |
| "caps_lambda_temporal": 0.0, | |
| "caps_noise_std": 0.05, | |
| "risk_distortion": "neutral", | |
| "risk_measure": 1.0, | |
| "device": "mps", | |
| "lr": 0.00044, | |
| "policy_lr": 0.00044, | |
| "log_every_updates": 200, | |
| "seed": 0, | |
| "log_dir": "runs/dsac_b747_step_response_20260106_021709" | |
| } | |
| } | |
| } |