{ "type": "act_romoya", "n_obs_steps": 1, "input_features": { "observation.state": { "type": "STATE", "shape": [ 7 ] }, "observation.images.wrist": { "type": "VISUAL", "shape": [ 3, 1080, 1920 ] }, "observation.images.top": { "type": "VISUAL", "shape": [ 3, 1080, 1920 ] }, "observation.images.side": { "type": "VISUAL", "shape": [ 3, 1080, 1920 ] } }, "output_features": { "action": { "type": "ACTION", "shape": [ 7 ] } }, "device": "cuda", "use_amp": false, "use_peft": false, "push_to_hub": true, "repo_id": "PL2011/act_sjaj_gripper-box", "private": null, "tags": null, "license": null, "pretrained_path": null, "chunk_size": 60, "n_action_steps": 5, "normalization_mapping": { "VISUAL": "MEAN_STD", "STATE": "MEAN_STD", "ACTION": "MEAN_STD" }, "vision_backbone": "resnet18", "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1", "replace_final_stride_with_dilation": false, "pre_norm": false, "dim_model": 512, "n_heads": 8, "dim_feedforward": 3200, "feedforward_activation": "relu", "n_encoder_layers": 4, "n_decoder_layers": 1, "use_vae": true, "latent_dim": 32, "n_vae_encoder_layers": 4, "temporal_ensemble_coeff": null, "dropout": 0.1, "kl_weight": 10.0, "optimizer_lr": 1e-05, "optimizer_weight_decay": 0.0001, "optimizer_lr_backbone": 1e-05, "action_mode": null, "state_feature_names_to_keep": [ "joint1.pos", "joint2.pos", "joint3.pos", "joint4.pos", "joint5.pos", "joint6.pos", "gripper.pos", "DO_1" ], "joint_action_names": [ "joint1.pos", "joint2.pos", "joint3.pos", "joint4.pos", "joint5.pos", "joint6.pos" ], "gripper_action_name": "gripper.pos", "do_action_names": [ "DO_1" ], "raw_observation_state_feature_names": [ "joint1.pos", "joint2.pos", "joint3.pos", "joint4.pos", "joint5.pos", "joint6.pos", "gripper.pos", "gripper.force", "DO_0", "DO_1", "joint1.effort", "joint2.effort", "joint3.effort", "joint4.effort", "joint5.effort", "joint6.effort", "joint1.vel", "joint2.vel", "joint3.vel", "joint4.vel", "joint5.vel", "joint6.vel", "joint1.acc", "joint2.acc", "joint3.acc", "joint4.acc", "joint5.acc", "joint6.acc", "tcp.x", "tcp.y", "tcp.z", "tcp.rx", "tcp.ry", "tcp.rz", "joint1.temp", "joint2.temp", "joint3.temp", "joint4.temp", "joint5.temp", "joint6.temp", "joint1.voltage", "joint2.voltage", "joint3.voltage", "joint4.voltage", "joint5.voltage", "joint6.voltage", "flange_voltage", "flange.x", "flange.y", "flange.z", "flange.rx", "flange.ry", "flange.rz", "target_joint1.pos", "target_joint2.pos", "target_joint3.pos", "target_joint4.pos", "target_joint5.pos", "target_joint6.pos", "target_joint1.vel", "target_joint2.vel", "target_joint3.vel", "target_joint4.vel", "target_joint5.vel", "target_joint6.vel", "target_joint1.acc", "target_joint2.acc", "target_joint3.acc", "target_joint4.acc", "target_joint5.acc", "target_joint6.acc", "target_joint1.effort", "target_joint2.effort", "target_joint3.effort", "target_joint4.effort", "target_joint5.effort", "target_joint6.effort", "target_tcp.x", "target_tcp.y", "target_tcp.z", "target_tcp.rx", "target_tcp.ry", "target_tcp.rz" ], "raw_action_feature_names": [ "joint1.pos", "joint2.pos", "joint3.pos", "joint4.pos", "joint5.pos", "joint6.pos", "gripper.pos", "gripper.force", "DO_0", "DO_1" ], "state_feature_names": [ "joint1.pos", "joint2.pos", "joint3.pos", "joint4.pos", "joint5.pos", "joint6.pos", "gripper.pos" ], "action_feature_names": [ "joint1.pos", "joint2.pos", "joint3.pos", "joint4.pos", "joint5.pos", "joint6.pos", "gripper.pos" ], "binary_state": [ null, null, null, null, null, null, null ], "binary_action": [ null, null, null, null, null, null, [ 50.0, 0.0, 99.0 ] ], "delta_action": [ false, false, false, false, false, false, false ], "do_threshold": 0.5, "sigmoid_do_outputs": false, "observation_image_resize_shape": [ 360, 640 ] }