{
    "type": "act_baseline",
    "n_obs_steps": 2,
    "input_features": {
        "observation.state": {
            "type": "STATE",
            "shape": [
                7
            ]
        },
        "observation.images.camera_0": {
            "type": "VISUAL",
            "shape": [
                3,
                480,
                640
            ]
        },
        "observation.images.camera_1": {
            "type": "VISUAL",
            "shape": [
                3,
                480,
                640
            ]
        }
    },
    "output_features": {
        "action": {
            "type": "ACTION",
            "shape": [
                7
            ]
        }
    },
    "device": "cuda",
    "use_amp": false,
    "use_peft": false,
    "push_to_hub": true,
    "repo_id": "IXDLI/act_baseline_pnp_long_100k",
    "private": null,
    "tags": null,
    "license": null,
    "pretrained_path": null,
    "horizon": 16,
    "n_action_steps": 8,
    "normalization_mapping": {
        "VISUAL": "MEAN_STD",
        "STATE": "MIN_MAX",
        "ACTION": "MIN_MAX"
    },
    "vision_backbone": "resnet18",
    "resize_shape": null,
    "crop_shape": [
        84,
        84
    ],
    "crop_is_random": true,
    "pretrained_backbone_weights": null,
    "use_group_norm": true,
    "spatial_softmax_num_keypoints": 32,
    "use_separate_rgb_encoder_per_camera": false,
    "hidden_dim": 512,
    "num_encoder_layers": 4,
    "num_decoder_layers": 1,
    "n_heads": 8,
    "dim_feedforward": 3200,
    "dropout": 0.1,
    "feedforward_activation": "relu",
    "pre_norm": false,
    "loss_type": "l1",
    "do_mask_loss_for_padding": true,
    "optimizer_lr": 1e-05,
    "optimizer_weight_decay": 0.0001,
    "optimizer_lr_backbone": 1e-05
}