train_config.json

{
    "dataset": {
        "repo_id": "lerobot/pusht",
        "episodes": null,
        "image_transforms": {
            "enable": false,
            "max_num_transforms": 3,
            "random_order": false,
            "tfs": {
                "brightness": {
                    "weight": 1.0,
                    "type": "ColorJitter",
                    "kwargs": {
                        "brightness": [
                            0.8,
                            1.2
                        ]
                    }
                },
                "contrast": {
                    "weight": 1.0,
                    "type": "ColorJitter",
                    "kwargs": {
                        "contrast": [
                            0.8,
                            1.2
                        ]
                    }
                },
                "saturation": {
                    "weight": 1.0,
                    "type": "ColorJitter",
                    "kwargs": {
                        "saturation": [
                            0.5,
                            1.5
                        ]
                    }
                },
                "hue": {
                    "weight": 1.0,
                    "type": "ColorJitter",
                    "kwargs": {
                        "hue": [
                            -0.05,
                            0.05
                        ]
                    }
                },
                "sharpness": {
                    "weight": 1.0,
                    "type": "SharpnessJitter",
                    "kwargs": {
                        "sharpness": [
                            0.5,
                            1.5
                        ]
                    }
                }
            }
        },
        "local_files_only": false,
        "use_imagenet_stats": true,
        "video_backend": "pyav"
    },
    "env": {
        "type": "pusht",
        "task": "PushT-v0",
        "fps": 10,
        "features": {
            "action": {
                "type": "ACTION",
                "shape": [
                    2
                ]
            },
            "agent_pos": {
                "type": "STATE",
                "shape": [
                    2
                ]
            },
            "pixels": {
                "type": "VISUAL",
                "shape": [
                    384,
                    384,
                    3
                ]
            }
        },
        "features_map": {
            "action": "action",
            "agent_pos": "observation.state",
            "environment_state": "observation.environment_state",
            "pixels": "observation.image"
        },
        "episode_length": 300,
        "obs_type": "pixels_agent_pos",
        "render_mode": "rgb_array",
        "visualization_width": 384,
        "visualization_height": 384
    },
    "policy": {
        "type": "dot",
        "n_obs_steps": 3,
        "normalization_mapping": {
            "VISUAL": "MEAN_STD",
            "STATE": "MIN_MAX",
            "ENV": "MIN_MAX",
            "ACTION": "MIN_MAX"
        },
        "input_features": {
            "observation.image": {
                "type": "VISUAL",
                "shape": [
                    3,
                    96,
                    96
                ]
            },
            "observation.state": {
                "type": "STATE",
                "shape": [
                    2
                ]
            }
        },
        "output_features": {
            "action": {
                "type": "ACTION",
                "shape": [
                    2
                ]
            }
        },
        "train_horizon": 20,
        "inference_horizon": 20,
        "lookback_obs_steps": 10,
        "lookback_aug": 5,
        "override_dataset_stats": false,
        "new_dataset_stats": {
            "action": {
                "max": [
                    512.0,
                    512.0
                ],
                "min": [
                    0.0,
                    0.0
                ]
            },
            "observation.environment_state": {
                "max": [
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0,
                    512.0
                ],
                "min": [
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0,
                    0.0
                ]
            },
            "observation.state": {
                "max": [
                    512.0,
                    512.0
                ],
                "min": [
                    0.0,
                    0.0
                ]
            }
        },
        "vision_backbone": "resnet18",
        "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
        "pre_norm": true,
        "lora_rank": 20,
        "merge_lora": false,
        "dim_model": 128,
        "n_heads": 8,
        "dim_feedforward": 512,
        "n_decoder_layers": 8,
        "rescale_shape": [
            96,
            96
        ],
        "crop_scale": 1.0,
        "state_noise": 0.01,
        "noise_decay": 0.999995,
        "dropout": 0.1,
        "alpha": 0.75,
        "train_alpha": 0.9,
        "predict_every_n": 1,
        "return_every_n": 2,
        "optimizer_lr": 0.0001,
        "optimizer_min_lr": 0.0001,
        "optimizer_lr_cycle_steps": 300000,
        "optimizer_weight_decay": 1e-05
    },
    "output_dir": "outputs/train/pusht_images",
    "job_name": "pusht_dot",
    "resume": false,
    "device": "cuda",
    "use_amp": true,
    "seed": 100000,
    "num_workers": 24,
    "batch_size": 24,
    "eval_freq": 10000,
    "log_freq": 1000,
    "save_checkpoint": true,
    "save_freq": 50000,
    "offline": {
          "steps": 1000000
      },
      "online": {
          "steps": 0,
          "rollout_n_episodes": 1,
          "rollout_batch_size": 1,
          "steps_between_rollouts": null,
          "sampling_ratio": 0.5,
          "env_seed": null,
          "buffer_capacity": null,
          "buffer_seed_size": 0,
          "do_rollout_async": false
      },
    "use_policy_training_preset": true,
    "optimizer": {
        "type": "adamw",
        "lr": 0.0001,
        "weight_decay": 1e-05,
        "grad_clip_norm": 10.0,
        "betas": [
            0.9,
            0.999
        ],
        "eps": 1e-08
    },
    "scheduler": {
        "type": "cosine_annealing",
        "num_warmup_steps": 0,
        "min_lr": 0.0001,
        "T_max": 300000
    },
    "eval": {
        "n_episodes": 50,
        "batch_size": 50,
        "use_async_envs": false
    },
    "wandb": {
        "enable": true,
        "disable_artifact": false,
        "project": "pusht",
        "entity": null,
        "notes": null
    }
}