| { | |
| "type": "uwm", | |
| "n_obs_steps": 1, | |
| "input_features": { | |
| "observation.state": { | |
| "type": "STATE", | |
| "shape": [ | |
| 6 | |
| ] | |
| }, | |
| "observation.images.wrist": { | |
| "type": "VISUAL", | |
| "shape": [ | |
| 3, | |
| 480, | |
| 640 | |
| ] | |
| }, | |
| "observation.images.exterior": { | |
| "type": "VISUAL", | |
| "shape": [ | |
| 3, | |
| 480, | |
| 640 | |
| ] | |
| } | |
| }, | |
| "output_features": { | |
| "action": { | |
| "type": "ACTION", | |
| "shape": [ | |
| 6 | |
| ] | |
| } | |
| }, | |
| "device": "cuda", | |
| "use_amp": false, | |
| "push_to_hub": true, | |
| "repo_id": "wszhaorobot/train_uwm_pick_one_object", | |
| "private": null, | |
| "tags": null, | |
| "license": null, | |
| "pretrained_path": null, | |
| "action_len": 100, | |
| "n_action_steps": 100, | |
| "normalization_mapping": { | |
| "VISUAL": "MEAN_STD", | |
| "STATE": "MIN_MAX", | |
| "ACTION": "MIN_MAX" | |
| }, | |
| "drop_n_last_frames": 0, | |
| "embed_dim": 768, | |
| "depth": 12, | |
| "num_heads": 12, | |
| "mlp_ratio": 4.0, | |
| "qkv_bias": true, | |
| "num_registers": 8, | |
| "timestep_embed_dim": 512, | |
| "latent_patch_shape": [ | |
| 1, | |
| 4, | |
| 4 | |
| ], | |
| "vision_backbone": "resnet", | |
| "resize_shape": [ | |
| 256, | |
| 256 | |
| ], | |
| "crop_shape": [ | |
| 224, | |
| 224 | |
| ], | |
| "random_crop": true, | |
| "imagenet_norm": false, | |
| "use_low_dim": true, | |
| "use_language": false, | |
| "num_train_timesteps": 100, | |
| "num_inference_steps": 10, | |
| "beta_schedule": "squaredcos_cap_v2", | |
| "clip_sample": true, | |
| "optimizer_lr": 0.0001, | |
| "optimizer_betas": [ | |
| 0.95, | |
| 0.999 | |
| ], | |
| "optimizer_eps": 1e-08, | |
| "optimizer_weight_decay": 1e-06, | |
| "scheduler_name": "cosine", | |
| "scheduler_warmup_steps": 500 | |
| } |