Hugging Face's logo

Dongkkka
/
ffw_sg2_rev1_PickCoffeePetEdit_SmolVLA_20k

Safetensors
Model card Files Community
Dongkkka commited on
Commit
ae05fae
·
verified ·
1 Parent(s): 1d58ddd

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. config.json +96 -0
  2. model.safetensors +3 -0
  3. train_config.json +208 -0
config.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.images.cam_head": {
11
+ "type": "VISUAL",
12
+ "shape": [
13
+ 3,
14
+ 376,
15
+ 672
16
+ ]
17
+ },
18
+ "observation.images.cam_wrist_left": {
19
+ "type": "VISUAL",
20
+ "shape": [
21
+ 3,
22
+ 240,
23
+ 424
24
+ ]
25
+ },
26
+ "observation.images.cam_wrist_right": {
27
+ "type": "VISUAL",
28
+ "shape": [
29
+ 3,
30
+ 240,
31
+ 424
32
+ ]
33
+ },
34
+ "observation.state": {
35
+ "type": "STATE",
36
+ "shape": [
37
+ 22
38
+ ]
39
+ }
40
+ },
41
+ "output_features": {
42
+ "action": {
43
+ "type": "ACTION",
44
+ "shape": [
45
+ 22
46
+ ]
47
+ }
48
+ },
49
+ "device": "cuda",
50
+ "use_amp": false,
51
+ "push_to_hub": true,
52
+ "repo_id": "Dongkkka/ffw_sg2_rev1_PickCoffeePetEdit_SmolVLA",
53
+ "private": null,
54
+ "tags": null,
55
+ "license": null,
56
+ "chunk_size": 50,
57
+ "n_action_steps": 50,
58
+ "max_state_dim": 32,
59
+ "max_action_dim": 32,
60
+ "resize_imgs_with_padding": [
61
+ 512,
62
+ 512
63
+ ],
64
+ "empty_cameras": 0,
65
+ "adapt_to_pi_aloha": false,
66
+ "use_delta_joint_actions_aloha": false,
67
+ "tokenizer_max_length": 48,
68
+ "num_steps": 10,
69
+ "use_cache": true,
70
+ "freeze_vision_encoder": true,
71
+ "train_expert_only": true,
72
+ "train_state_proj": true,
73
+ "optimizer_lr": 0.0001,
74
+ "optimizer_betas": [
75
+ 0.9,
76
+ 0.95
77
+ ],
78
+ "optimizer_eps": 1e-08,
79
+ "optimizer_weight_decay": 1e-10,
80
+ "optimizer_grad_clip_norm": 10.0,
81
+ "scheduler_warmup_steps": 1000,
82
+ "scheduler_decay_steps": 30000,
83
+ "scheduler_decay_lr": 2.5e-06,
84
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
85
+ "load_vlm_weights": true,
86
+ "add_image_special_tokens": false,
87
+ "attention_mode": "cross_attn",
88
+ "prefix_length": 0,
89
+ "pad_language_to": "max_length",
90
+ "num_expert_layers": 0,
91
+ "num_vlm_layers": 16,
92
+ "self_attn_every_n_layers": 2,
93
+ "expert_width_multiplier": 0.75,
94
+ "min_period": 0.004,
95
+ "max_period": 4.0
96
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6563a61f62f02ae69124af0bc661029e3237c61002ab9bb3275cc08f11cfbd4f
3
+ size 906713688
train_config.json ADDED
@@ -0,0 +1,208 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "Dongkkka/ffw_sg2_rev1_PickCoffeePetEdit",
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
+ },
67
+ "env": null,
68
+ "policy": {
69
+ "type": "smolvla",
70
+ "n_obs_steps": 1,
71
+ "normalization_mapping": {
72
+ "VISUAL": "IDENTITY",
73
+ "STATE": "MEAN_STD",
74
+ "ACTION": "MEAN_STD"
75
+ },
76
+ "input_features": {
77
+ "observation.images.cam_head": {
78
+ "type": "VISUAL",
79
+ "shape": [
80
+ 3,
81
+ 376,
82
+ 672
83
+ ]
84
+ },
85
+ "observation.images.cam_wrist_left": {
86
+ "type": "VISUAL",
87
+ "shape": [
88
+ 3,
89
+ 240,
90
+ 424
91
+ ]
92
+ },
93
+ "observation.images.cam_wrist_right": {
94
+ "type": "VISUAL",
95
+ "shape": [
96
+ 3,
97
+ 240,
98
+ 424
99
+ ]
100
+ },
101
+ "observation.state": {
102
+ "type": "STATE",
103
+ "shape": [
104
+ 22
105
+ ]
106
+ }
107
+ },
108
+ "output_features": {
109
+ "action": {
110
+ "type": "ACTION",
111
+ "shape": [
112
+ 22
113
+ ]
114
+ }
115
+ },
116
+ "device": "cuda",
117
+ "use_amp": false,
118
+ "push_to_hub": true,
119
+ "repo_id": "Dongkkka/ffw_sg2_rev1_PickCoffeePetEdit_SmolVLA",
120
+ "private": null,
121
+ "tags": null,
122
+ "license": null,
123
+ "chunk_size": 50,
124
+ "n_action_steps": 50,
125
+ "max_state_dim": 32,
126
+ "max_action_dim": 32,
127
+ "resize_imgs_with_padding": [
128
+ 512,
129
+ 512
130
+ ],
131
+ "empty_cameras": 0,
132
+ "adapt_to_pi_aloha": false,
133
+ "use_delta_joint_actions_aloha": false,
134
+ "tokenizer_max_length": 48,
135
+ "num_steps": 10,
136
+ "use_cache": true,
137
+ "freeze_vision_encoder": true,
138
+ "train_expert_only": true,
139
+ "train_state_proj": true,
140
+ "optimizer_lr": 0.0001,
141
+ "optimizer_betas": [
142
+ 0.9,
143
+ 0.95
144
+ ],
145
+ "optimizer_eps": 1e-08,
146
+ "optimizer_weight_decay": 1e-10,
147
+ "optimizer_grad_clip_norm": 10.0,
148
+ "scheduler_warmup_steps": 1000,
149
+ "scheduler_decay_steps": 30000,
150
+ "scheduler_decay_lr": 2.5e-06,
151
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
152
+ "load_vlm_weights": true,
153
+ "add_image_special_tokens": false,
154
+ "attention_mode": "cross_attn",
155
+ "prefix_length": 0,
156
+ "pad_language_to": "max_length",
157
+ "num_expert_layers": 0,
158
+ "num_vlm_layers": 16,
159
+ "self_attn_every_n_layers": 2,
160
+ "expert_width_multiplier": 0.75,
161
+ "min_period": 0.004,
162
+ "max_period": 4.0
163
+ },
164
+ "output_dir": "outputs/train/ffw_sg2_rev1_PickCoffeePetEdit_SmolVLA",
165
+ "job_name": "my_smolvla_training",
166
+ "resume": false,
167
+ "seed": 1000,
168
+ "num_workers": 16,
169
+ "batch_size": 10,
170
+ "steps": 100000,
171
+ "eval_freq": 5000,
172
+ "log_freq": 200,
173
+ "save_checkpoint": true,
174
+ "save_freq": 5000,
175
+ "use_policy_training_preset": true,
176
+ "optimizer": {
177
+ "type": "adamw",
178
+ "lr": 0.0001,
179
+ "weight_decay": 1e-10,
180
+ "grad_clip_norm": 10.0,
181
+ "betas": [
182
+ 0.9,
183
+ 0.95
184
+ ],
185
+ "eps": 1e-08
186
+ },
187
+ "scheduler": {
188
+ "type": "cosine_decay_with_warmup",
189
+ "num_warmup_steps": 1000,
190
+ "num_decay_steps": 30000,
191
+ "peak_lr": 0.0001,
192
+ "decay_lr": 2.5e-06
193
+ },
194
+ "eval": {
195
+ "n_episodes": 50,
196
+ "batch_size": 50,
197
+ "use_async_envs": false
198
+ },
199
+ "wandb": {
200
+ "enable": false,
201
+ "disable_artifact": false,
202
+ "project": "lerobot",
203
+ "entity": null,
204
+ "notes": null,
205
+ "run_id": null,
206
+ "mode": null
207
+ }
208
+ }