Hugging Face's logo

ChihHanShen
/
multitask-dit-so101

Robotics
LeRobot
Safetensors
multi_task_dit
Model card Files
xet
 Community
ChihHanShen commited on
Commit
d4a1e0f
·
verified ·
1 Parent(s): 41b291a

Upload policy

Browse files
Files changed (2) hide show
  1. config.json +114 -0
  2. model.safetensors +3 -0
config.json ADDED
@@ -0,0 +1,114 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "multi_task_dit",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 12
9
+ ]
10
+ },
11
+ "observation.images.wrist_left": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 480,
16
+ 640
17
+ ]
18
+ },
19
+ "observation.images.center": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 480,
24
+ 640
25
+ ]
26
+ },
27
+ "observation.images.wrist_right": {
28
+ "type": "VISUAL",
29
+ "shape": [
30
+ 3,
31
+ 480,
32
+ 640
33
+ ]
34
+ }
35
+ },
36
+ "output_features": {
37
+ "action": {
38
+ "type": "ACTION",
39
+ "shape": [
40
+ 12
41
+ ]
42
+ }
43
+ },
44
+ "device": "cuda",
45
+ "use_amp": true,
46
+ "use_peft": false,
47
+ "push_to_hub": true,
48
+ "repo_id": "ChihHanShen/multitask-dit-so101",
49
+ "private": null,
50
+ "tags": null,
51
+ "license": null,
52
+ "pretrained_path": null,
53
+ "pretrained_revision": null,
54
+ "horizon": 32,
55
+ "n_action_steps": 24,
56
+ "objective": "diffusion",
57
+ "noise_scheduler_type": "DDPM",
58
+ "num_train_timesteps": 100,
59
+ "beta_schedule": "squaredcos_cap_v2",
60
+ "beta_start": 0.0001,
61
+ "beta_end": 0.02,
62
+ "prediction_type": "epsilon",
63
+ "clip_sample": true,
64
+ "clip_sample_range": 1.0,
65
+ "num_inference_steps": null,
66
+ "sigma_min": 0.0,
67
+ "num_integration_steps": 100,
68
+ "integration_method": "euler",
69
+ "timestep_sampling_strategy": "beta",
70
+ "timestep_sampling_s": 0.999,
71
+ "timestep_sampling_alpha": 1.5,
72
+ "timestep_sampling_beta": 1.0,
73
+ "hidden_dim": 512,
74
+ "num_layers": 6,
75
+ "num_heads": 8,
76
+ "dropout": 0.1,
77
+ "use_positional_encoding": false,
78
+ "timestep_embed_dim": 256,
79
+ "use_rope": true,
80
+ "rope_base": 10000.0,
81
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
82
+ "use_separate_rgb_encoder_per_camera": false,
83
+ "vision_encoder_lr_multiplier": 0.1,
84
+ "image_resize_shape": [
85
+ 256,
86
+ 256
87
+ ],
88
+ "image_crop_shape": [
89
+ 224,
90
+ 224
91
+ ],
92
+ "image_crop_is_random": true,
93
+ "text_encoder_name": "openai/clip-vit-base-patch16",
94
+ "tokenizer_max_length": 77,
95
+ "tokenizer_padding": "max_length",
96
+ "tokenizer_padding_side": "right",
97
+ "tokenizer_truncation": true,
98
+ "normalization_mapping": {
99
+ "VISUAL": "MEAN_STD",
100
+ "STATE": "MIN_MAX",
101
+ "ACTION": "MIN_MAX"
102
+ },
103
+ "optimizer_lr": 0.0001,
104
+ "optimizer_betas": [
105
+ 0.95,
106
+ 0.999
107
+ ],
108
+ "optimizer_eps": 1e-08,
109
+ "optimizer_weight_decay": 0.0,
110
+ "scheduler_name": "cosine",
111
+ "scheduler_warmup_steps": 2000,
112
+ "do_mask_loss_for_padding": false,
113
+ "drop_n_last_frames": 7
114
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68838e08a39a77bb4fd407c9676224846b3ee145c776b306d1b5dde13deafd1e
3
+ size 1109649816