Songyuan666/robot_learning_tutorial_diffusion_example_model

1{
2    "type": "diffusion",
3    "n_obs_steps": 2,
4    "input_features": {
5        "observation.state": {
6            "type": "STATE",
7            "shape": [
8                6
9            ]
10        },
11        "observation.images.side": {
12            "type": "VISUAL",
13            "shape": [
14                3,
15                480,
16                640
17            ]
18        },
19        "observation.images.front": {
20            "type": "VISUAL",
21            "shape": [
22                3,
23                480,
24                640
25            ]
26        }
27    },
28    "output_features": {
29        "action": {
30            "type": "ACTION",
31            "shape": [
32                6
33            ]
34        }
35    },
36    "device": "cuda",
37    "use_amp": false,
38    "use_peft": false,
39    "push_to_hub": true,
40    "repo_id": null,
41    "private": null,
42    "tags": null,
43    "license": null,
44    "pretrained_path": null,
45    "horizon": 16,
46    "n_action_steps": 8,
47    "normalization_mapping": {
48        "VISUAL": "MEAN_STD",
49        "STATE": "MIN_MAX",
50        "ACTION": "MIN_MAX"
51    },
52    "drop_n_last_frames": 7,
53    "vision_backbone": "resnet18",
54    "resize_shape": null,
55    "crop_ratio": 1.0,
56    "crop_shape": null,
57    "crop_is_random": true,
58    "pretrained_backbone_weights": null,
59    "use_group_norm": true,
60    "spatial_softmax_num_keypoints": 32,
61    "use_separate_rgb_encoder_per_camera": false,
62    "down_dims": [
63        512,
64        1024,
65        2048
66    ],
67    "kernel_size": 5,
68    "n_groups": 8,
69    "diffusion_step_embed_dim": 128,
70    "use_film_scale_modulation": true,
71    "noise_scheduler_type": "DDPM",
72    "num_train_timesteps": 100,
73    "beta_schedule": "squaredcos_cap_v2",
74    "beta_start": 0.0001,
75    "beta_end": 0.02,
76    "prediction_type": "epsilon",
77    "clip_sample": true,
78    "clip_sample_range": 1.0,
79    "num_inference_steps": null,
80    "compile_model": false,
81    "compile_mode": "reduce-overhead",
82    "do_mask_loss_for_padding": false,
83    "optimizer_lr": 0.0001,
84    "optimizer_betas": [
85        0.95,
86        0.999
87    ],
88    "optimizer_eps": 1e-08,
89    "optimizer_weight_decay": 1e-06,
90    "scheduler_name": "cosine",
91    "scheduler_warmup_steps": 500
92}