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": "vqbet",
        "n_obs_steps": 5,
        "normalization_mapping": {
            "VISUAL": "IDENTITY",
            "STATE": "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
                ]
            }
        },
        "n_action_pred_token": 3,
        "action_chunk_size": 5,
        "vision_backbone": "resnet18",
        "crop_shape": [
            84,
            84
        ],
        "crop_is_random": true,
        "pretrained_backbone_weights": null,
        "use_group_norm": true,
        "spatial_softmax_num_keypoints": 32,
        "n_vqvae_training_steps": 20000,
        "vqvae_n_embed": 16,
        "vqvae_embedding_dim": 256,
        "vqvae_enc_hidden_dim": 128,
        "gpt_block_size": 500,
        "gpt_input_dim": 512,
        "gpt_output_dim": 512,
        "gpt_n_layer": 8,
        "gpt_n_head": 8,
        "gpt_hidden_dim": 512,
        "dropout": 0.1,
        "mlp_hidden_dim": 1024,
        "offset_loss_weight": 10000.0,
        "primary_code_loss_weight": 5.0,
        "secondary_code_loss_weight": 0.5,
        "bet_softmax_temperature": 0.1,
        "sequentially_select": false,
        "optimizer_lr": 0.0001,
        "optimizer_betas": [
            0.95,
            0.999
        ],
        "optimizer_eps": 1e-08,
        "optimizer_weight_decay": 1e-06,
        "optimizer_vqvae_lr": 0.001,
        "optimizer_vqvae_weight_decay": 0.0001,
        "scheduler_warmup_steps": 500
    },
    "output_dir": "outputs/train/2025-01-25/20-30-40_pusht_vqbet",
    "job_name": "pusht_vqbet",
    "resume": false,
    "device": "cuda",
    "use_amp": false,
    "seed": 100000,
    "num_workers": 4,
    "batch_size": 64,
    "steps": 250000,
    "eval_freq": 25000,
    "log_freq": 200,
    "save_checkpoint": true,
    "save_freq": 25000,
    "use_policy_training_preset": true,
    "optimizer": {
        "type": "adam",
        "lr": 0.0001,
        "betas": [
            0.95,
            0.999
        ],
        "eps": 1e-08,
        "weight_decay": 1e-06,
        "grad_clip_norm": 10.0
    },
    "scheduler": {
        "type": "vqbet",
        "num_warmup_steps": 500,
        "num_vqvae_training_steps": 20000,
        "num_cycles": 0.5
    },
    "eval": {
        "n_episodes": 50,
        "batch_size": 50,
        "use_async_envs": false
    },
    "wandb": {
        "enable": true,
        "disable_artifact": false,
        "project": "lerobot",
        "entity": null,
        "notes": null
    }
}