Hugging Face's logo

satvikahuja
/
cube_smolvla_regular_data_20k

Safetensors
Model card Files
xet
Community
satvikahuja commited on
Commit
b1d8c5b
·
verified ·
1 Parent(s): e8fd6c8

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. config.json +88 -0
  2. model.safetensors +3 -0
  3. train_config.json +201 -0
config.json ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "smolvla",
3
+ "n_obs_steps": 1,
4
+ "normalization_mapping": {
5
+ "VISUAL": "IDENTITY",
6
+ "STATE": "MEAN_STD",
7
+ "ACTION": "MEAN_STD"
8
+ },
9
+ "input_features": {
10
+ "observation.state": {
11
+ "type": "STATE",
12
+ "shape": [
13
+ 6
14
+ ]
15
+ },
16
+ "observation.images.gripper": {
17
+ "type": "VISUAL",
18
+ "shape": [
19
+ 3,
20
+ 480,
21
+ 640
22
+ ]
23
+ },
24
+ "observation.images.top": {
25
+ "type": "VISUAL",
26
+ "shape": [
27
+ 3,
28
+ 480,
29
+ 640
30
+ ]
31
+ }
32
+ },
33
+ "output_features": {
34
+ "action": {
35
+ "type": "ACTION",
36
+ "shape": [
37
+ 6
38
+ ]
39
+ }
40
+ },
41
+ "device": "cuda",
42
+ "use_amp": false,
43
+ "push_to_hub": true,
44
+ "repo_id": "satvikahuja/cube_regular_data",
45
+ "private": null,
46
+ "tags": null,
47
+ "license": null,
48
+ "chunk_size": 50,
49
+ "n_action_steps": 50,
50
+ "max_state_dim": 32,
51
+ "max_action_dim": 32,
52
+ "resize_imgs_with_padding": [
53
+ 512,
54
+ 512
55
+ ],
56
+ "empty_cameras": 0,
57
+ "adapt_to_pi_aloha": false,
58
+ "use_delta_joint_actions_aloha": false,
59
+ "tokenizer_max_length": 48,
60
+ "num_steps": 10,
61
+ "use_cache": true,
62
+ "freeze_vision_encoder": true,
63
+ "train_expert_only": true,
64
+ "train_state_proj": true,
65
+ "optimizer_lr": 0.0001,
66
+ "optimizer_betas": [
67
+ 0.9,
68
+ 0.95
69
+ ],
70
+ "optimizer_eps": 1e-08,
71
+ "optimizer_weight_decay": 1e-10,
72
+ "optimizer_grad_clip_norm": 10.0,
73
+ "scheduler_warmup_steps": 1000,
74
+ "scheduler_decay_steps": 30000,
75
+ "scheduler_decay_lr": 2.5e-06,
76
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
77
+ "load_vlm_weights": true,
78
+ "add_image_special_tokens": false,
79
+ "attention_mode": "cross_attn",
80
+ "prefix_length": 0,
81
+ "pad_language_to": "max_length",
82
+ "num_expert_layers": 0,
83
+ "num_vlm_layers": 16,
84
+ "self_attn_every_n_layers": 2,
85
+ "expert_width_multiplier": 0.75,
86
+ "min_period": 0.004,
87
+ "max_period": 4.0
88
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:05bfff1b80840937455503f821e58eededf9f578da70b572ce1f5eee95ef4804
3
+ size 906713296
train_config.json ADDED
@@ -0,0 +1,201 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "satvikahuja/cube_regular_data",
4
+ "root": null,
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": false,
8
+ "max_num_transforms": 3,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "brightness": {
12
+ "weight": 1.0,
13
+ "type": "ColorJitter",
14
+ "kwargs": {
15
+ "brightness": [
16
+ 0.8,
17
+ 1.2
18
+ ]
19
+ }
20
+ },
21
+ "contrast": {
22
+ "weight": 1.0,
23
+ "type": "ColorJitter",
24
+ "kwargs": {
25
+ "contrast": [
26
+ 0.8,
27
+ 1.2
28
+ ]
29
+ }
30
+ },
31
+ "saturation": {
32
+ "weight": 1.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "saturation": [
36
+ 0.5,
37
+ 1.5
38
+ ]
39
+ }
40
+ },
41
+ "hue": {
42
+ "weight": 1.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "hue": [
46
+ -0.05,
47
+ 0.05
48
+ ]
49
+ }
50
+ },
51
+ "sharpness": {
52
+ "weight": 1.0,
53
+ "type": "SharpnessJitter",
54
+ "kwargs": {
55
+ "sharpness": [
56
+ 0.5,
57
+ 1.5
58
+ ]
59
+ }
60
+ }
61
+ }
62
+ },
63
+ "revision": null,
64
+ "use_imagenet_stats": true,
65
+ "video_backend": "torchcodec",
66
+ "streaming": false
67
+ },
68
+ "env": null,
69
+ "policy": {
70
+ "type": "smolvla",
71
+ "n_obs_steps": 1,
72
+ "normalization_mapping": {
73
+ "VISUAL": "IDENTITY",
74
+ "STATE": "MEAN_STD",
75
+ "ACTION": "MEAN_STD"
76
+ },
77
+ "input_features": {
78
+ "observation.state": {
79
+ "type": "STATE",
80
+ "shape": [
81
+ 6
82
+ ]
83
+ },
84
+ "observation.images.gripper": {
85
+ "type": "VISUAL",
86
+ "shape": [
87
+ 3,
88
+ 480,
89
+ 640
90
+ ]
91
+ },
92
+ "observation.images.top": {
93
+ "type": "VISUAL",
94
+ "shape": [
95
+ 3,
96
+ 480,
97
+ 640
98
+ ]
99
+ }
100
+ },
101
+ "output_features": {
102
+ "action": {
103
+ "type": "ACTION",
104
+ "shape": [
105
+ 6
106
+ ]
107
+ }
108
+ },
109
+ "device": "cuda",
110
+ "use_amp": false,
111
+ "push_to_hub": true,
112
+ "repo_id": "satvikahuja/cube_regular_data",
113
+ "private": null,
114
+ "tags": null,
115
+ "license": null,
116
+ "chunk_size": 50,
117
+ "n_action_steps": 50,
118
+ "max_state_dim": 32,
119
+ "max_action_dim": 32,
120
+ "resize_imgs_with_padding": [
121
+ 512,
122
+ 512
123
+ ],
124
+ "empty_cameras": 0,
125
+ "adapt_to_pi_aloha": false,
126
+ "use_delta_joint_actions_aloha": false,
127
+ "tokenizer_max_length": 48,
128
+ "num_steps": 10,
129
+ "use_cache": true,
130
+ "freeze_vision_encoder": true,
131
+ "train_expert_only": true,
132
+ "train_state_proj": true,
133
+ "optimizer_lr": 0.0001,
134
+ "optimizer_betas": [
135
+ 0.9,
136
+ 0.95
137
+ ],
138
+ "optimizer_eps": 1e-08,
139
+ "optimizer_weight_decay": 1e-10,
140
+ "optimizer_grad_clip_norm": 10.0,
141
+ "scheduler_warmup_steps": 1000,
142
+ "scheduler_decay_steps": 30000,
143
+ "scheduler_decay_lr": 2.5e-06,
144
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
145
+ "load_vlm_weights": true,
146
+ "add_image_special_tokens": false,
147
+ "attention_mode": "cross_attn",
148
+ "prefix_length": 0,
149
+ "pad_language_to": "max_length",
150
+ "num_expert_layers": 0,
151
+ "num_vlm_layers": 16,
152
+ "self_attn_every_n_layers": 2,
153
+ "expert_width_multiplier": 0.75,
154
+ "min_period": 0.004,
155
+ "max_period": 4.0
156
+ },
157
+ "output_dir": "outputs/train/cube_regular_data",
158
+ "job_name": "cube_regular_data",
159
+ "resume": false,
160
+ "seed": 1000,
161
+ "num_workers": 4,
162
+ "batch_size": 64,
163
+ "steps": 200000,
164
+ "eval_freq": 5000,
165
+ "log_freq": 200,
166
+ "save_checkpoint": true,
167
+ "save_freq": 5000,
168
+ "use_policy_training_preset": true,
169
+ "optimizer": {
170
+ "type": "adamw",
171
+ "lr": 0.0001,
172
+ "weight_decay": 1e-10,
173
+ "grad_clip_norm": 10.0,
174
+ "betas": [
175
+ 0.9,
176
+ 0.95
177
+ ],
178
+ "eps": 1e-08
179
+ },
180
+ "scheduler": {
181
+ "type": "cosine_decay_with_warmup",
182
+ "num_warmup_steps": 1000,
183
+ "num_decay_steps": 30000,
184
+ "peak_lr": 0.0001,
185
+ "decay_lr": 2.5e-06
186
+ },
187
+ "eval": {
188
+ "n_episodes": 50,
189
+ "batch_size": 50,
190
+ "use_async_envs": false
191
+ },
192
+ "wandb": {
193
+ "enable": true,
194
+ "disable_artifact": false,
195
+ "project": "lerobot",
196
+ "entity": null,
197
+ "notes": null,
198
+ "run_id": "v0zd8uf8",
199
+ "mode": null
200
+ }
201
+ }