| { |
| "type": "lingbot_va", |
| "n_obs_steps": 1, |
| "input_features": { |
| "observation.images.cam_high": { |
| "type": "VISUAL", |
| "shape": [ |
| 3, |
| 256, |
| 256 |
| ] |
| }, |
| "observation.images.cam_left_wrist": { |
| "type": "VISUAL", |
| "shape": [ |
| 3, |
| 256, |
| 256 |
| ] |
| }, |
| "observation.images.cam_right_wrist": { |
| "type": "VISUAL", |
| "shape": [ |
| 3, |
| 256, |
| 256 |
| ] |
| } |
| }, |
| "output_features": { |
| "action": { |
| "type": "ACTION", |
| "shape": [ |
| 16 |
| ] |
| } |
| }, |
| "device": "cuda", |
| "use_amp": false, |
| "use_peft": false, |
| "push_to_hub": true, |
| "repo_id": "maximellerbach/folding_lingbot_epanchor", |
| "private": null, |
| "tags": null, |
| "license": null, |
| "pretrained_path": "/fsx/maxime_ellerbach/robots/runs/folding-tshirt/outputs/train/lingbot_absolute_22372322/checkpoints/last/pretrained_model", |
| "pretrained_revision": null, |
| "patch_size": [ |
| 1, |
| 2, |
| 2 |
| ], |
| "num_attention_heads": 24, |
| "attention_head_dim": 128, |
| "in_channels": 48, |
| "out_channels": 48, |
| "action_dim": 30, |
| "text_dim": 4096, |
| "freq_dim": 256, |
| "ffn_dim": 14336, |
| "num_layers": 30, |
| "cross_attn_norm": true, |
| "eps": 1e-06, |
| "rope_max_seq_len": 1024, |
| "attn_mode": "flex", |
| "wan_pretrained_path": "robbyant/lingbot-va-base", |
| "dtype": "bfloat16", |
| "text_encoder_device": "cuda", |
| "obs_cam_keys": [ |
| "observation.images.cam_high", |
| "observation.images.cam_left_wrist", |
| "observation.images.cam_right_wrist" |
| ], |
| "image_hflip": false, |
| "camera_layout": "width_concat", |
| "height": 224, |
| "width": 224, |
| "action_per_frame": 16, |
| "frame_chunk_size": 2, |
| "attn_window": 72, |
| "num_inference_steps": 25, |
| "video_exec_step": -1, |
| "action_num_inference_steps": 50, |
| "guidance_scale": 5.0, |
| "action_guidance_scale": 1.0, |
| "snr_shift": 5.0, |
| "action_snr_shift": 1.0, |
| "max_sequence_length": 512, |
| "used_action_channel_ids": [ |
| 21, |
| 22, |
| 23, |
| 24, |
| 25, |
| 26, |
| 27, |
| 29, |
| 14, |
| 15, |
| 16, |
| 17, |
| 18, |
| 19, |
| 20, |
| 28 |
| ], |
| "action_anchor": "episode", |
| "action_anchor_exclude_joints": [ |
| "gripper" |
| ], |
| "action_feature_names": [ |
| "right_joint_1.pos", |
| "right_joint_2.pos", |
| "right_joint_3.pos", |
| "right_joint_4.pos", |
| "right_joint_5.pos", |
| "right_joint_6.pos", |
| "right_joint_7.pos", |
| "right_gripper.pos", |
| "left_joint_1.pos", |
| "left_joint_2.pos", |
| "left_joint_3.pos", |
| "left_joint_4.pos", |
| "left_joint_5.pos", |
| "left_joint_6.pos", |
| "left_joint_7.pos", |
| "left_gripper.pos" |
| ], |
| "train_latent_frames": 8, |
| "save_predicted_video": false, |
| "normalization_mapping": { |
| "VISUAL": "IDENTITY", |
| "STATE": "IDENTITY", |
| "ACTION": "QUANTILES" |
| }, |
| "optimizer_lr": 2e-05, |
| "optimizer_betas": [ |
| 0.9, |
| 0.95 |
| ], |
| "optimizer_eps": 1e-08, |
| "optimizer_weight_decay": 0.0001, |
| "optimizer_grad_clip_norm": 1.0, |
| "scheduler_warmup_steps": 1000, |
| "scheduler_type": "constant_with_warmup", |
| "action_noisy_cond_prob": 0.0 |
| } |
