| { | |
| "dataset": { | |
| "repo_id": "useless", | |
| "root": "/home/ps/ssd/Data/X-VLA/Dataset/1203_pm_softfold_compressed_lerobot", | |
| "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 | |
| ] | |
| } | |
| }, | |
| "affine": { | |
| "weight": 1.0, | |
| "type": "RandomAffine", | |
| "kwargs": { | |
| "degrees": [ | |
| -5.0, | |
| 5.0 | |
| ], | |
| "translate": [ | |
| 0.05, | |
| 0.05 | |
| ] | |
| } | |
| } | |
| } | |
| }, | |
| "revision": null, | |
| "use_imagenet_stats": true, | |
| "video_backend": "torchcodec", | |
| "streaming": false | |
| }, | |
| "env": null, | |
| "policy": { | |
| "type": "sarm", | |
| "n_obs_steps": 8, | |
| "input_features": { | |
| "observation.images.cam_high": { | |
| "type": "VISUAL", | |
| "shape": [ | |
| 480, | |
| 640, | |
| 3 | |
| ] | |
| }, | |
| "observation.state": { | |
| "type": "STATE", | |
| "shape": [ | |
| 32 | |
| ] | |
| } | |
| }, | |
| "output_features": { | |
| "action": { | |
| "type": "ACTION", | |
| "shape": [ | |
| 14 | |
| ] | |
| } | |
| }, | |
| "device": "cuda", | |
| "use_amp": false, | |
| "push_to_hub": true, | |
| "repo_id": "useless", | |
| "private": null, | |
| "tags": null, | |
| "license": null, | |
| "pretrained_path": null, | |
| "annotation_mode": "dual", | |
| "frame_gap": 30, | |
| "max_rewind_steps": 4, | |
| "image_dim": 512, | |
| "text_dim": 512, | |
| "hidden_dim": 768, | |
| "num_heads": 12, | |
| "num_layers": 8, | |
| "max_state_dim": 32, | |
| "drop_n_last_frames": 1, | |
| "batch_size": 64, | |
| "clip_batch_size": 64, | |
| "dropout": 0.1, | |
| "stage_loss_weight": 1.0, | |
| "rewind_probability": 0.8, | |
| "language_perturbation_probability": 0.2, | |
| "num_sparse_stages": 3, | |
| "sparse_subtask_names": [ | |
| "Flatten the tshirt out", | |
| "fold the tshirt", | |
| "put folded tshirt to the right side" | |
| ], | |
| "sparse_temporal_proportions": [ | |
| 0.42229928493207264, | |
| 0.4390527818761901, | |
| 0.13864793319173732 | |
| ], | |
| "num_dense_stages": 8, | |
| "dense_subtask_names": [ | |
| "Move arms from starting position to the workspace", | |
| "manipulate the fabric to reveal the hem", | |
| "grasp the bottom corners and flick forward to flatten the tshirt", | |
| "grasp the armpit areas and fold the upper half forward", | |
| "fold the sleeves inward", | |
| "grasp the bottom hem to fold the tshirt vertically", | |
| "grasp the right side to fold it horizontally", | |
| "pick up the folded tshirt and move it to the right side" | |
| ], | |
| "dense_temporal_proportions": [ | |
| 0.04753523919499561, | |
| 0.27640619776466735, | |
| 0.10446237491517536, | |
| 0.09002541761368966, | |
| 0.12282106557613906, | |
| 0.13770028217609287, | |
| 0.08866373242394851, | |
| 0.13238569033529157 | |
| ], | |
| "pretrained_model_path": null, | |
| "image_key": "observation.images.cam_high", | |
| "state_key": "observation.state", | |
| "normalization_mapping": { | |
| "VISUAL": "IDENTITY", | |
| "STATE": "MEAN_STD", | |
| "LANGUAGE": "IDENTITY", | |
| "REWARD": "IDENTITY" | |
| } | |
| }, | |
| "output_dir": "/media/ps/My_Passport/Data/SARM/outputs/train/sarm_dual_50k_iters", | |
| "job_name": "sarm", | |
| "resume": false, | |
| "seed": 1000, | |
| "num_workers": 4, | |
| "batch_size": 32, | |
| "steps": 50000, | |
| "eval_freq": 20000, | |
| "log_freq": 200, | |
| "tolerance_s": 0.0001, | |
| "save_checkpoint": true, | |
| "save_freq": 20000, | |
| "use_policy_training_preset": true, | |
| "optimizer": { | |
| "type": "adamw", | |
| "lr": 5e-05, | |
| "weight_decay": 0.001, | |
| "grad_clip_norm": 10.0, | |
| "betas": [ | |
| 0.9, | |
| 0.999 | |
| ], | |
| "eps": 1e-08 | |
| }, | |
| "scheduler": { | |
| "type": "cosine_decay_with_warmup", | |
| "num_warmup_steps": 500, | |
| "num_decay_steps": 50000, | |
| "peak_lr": 5e-05, | |
| "decay_lr": 5e-06 | |
| }, | |
| "eval": { | |
| "n_episodes": 50, | |
| "batch_size": 50, | |
| "use_async_envs": false | |
| }, | |
| "wandb": { | |
| "enable": true, | |
| "disable_artifact": false, | |
| "project": "sarm", | |
| "entity": null, | |
| "notes": null, | |
| "run_id": "wmue4whw", | |
| "mode": null | |
| }, | |
| "use_rabc": false, | |
| "rabc_progress_path": null, | |
| "rabc_kappa": 0.01, | |
| "rabc_epsilon": 1e-06, | |
| "rabc_head_mode": "sparse", | |
| "rename_map": {}, | |
| "checkpoint_path": null | |
| } |