| { | |
| "action_dim": 8, | |
| "action_horizon": 4, | |
| "batch_groups": 2, | |
| "batch_size_groups": null, | |
| "dataset_dir": "outputs/phase5_train_debug/cil", | |
| "device": "auto", | |
| "effect_dim": 32, | |
| "epochs": 1, | |
| "hidden_dim": 64, | |
| "lang_dim": 64, | |
| "learning_rate": 0.001, | |
| "losses": { | |
| "bc": 1.0, | |
| "bc_best_action": 1.0, | |
| "causal_contrastive": 0.5, | |
| "contrast": 0.5, | |
| "effect": 1.0, | |
| "forward_effect_prediction": 1.0, | |
| "lang_pair": 0.25, | |
| "language_minimal_pair": 0.25, | |
| "progress": 1.0, | |
| "rank": 1.0, | |
| "regret": 0.5, | |
| "regret_prediction": 0.5, | |
| "same_state_pairwise_ranking": 1.0, | |
| "success": 1.0 | |
| }, | |
| "lr": null, | |
| "obs_dim": 32, | |
| "output_dir": "outputs/phase5_train_debug/run", | |
| "pair_count_per_group": 8, | |
| "records_per_group": 4, | |
| "seed": 0, | |
| "val_fraction": 0.2, | |
| "wandb": false, | |
| "weight_decay": 0.0 | |
| } | |