Hugging Face's logo

continuallearning
/
smolvla_image_only_plate_base_lsy_stack

Safetensors
Model card Files
xet
Community
continuallearning commited on
Commit
40149e4
·
verified ·
1 Parent(s): a25c613

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. config.json +90 -0
  2. model.safetensors +3 -0
  3. train_config.json +204 -0
config.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "smolvla_image_only",
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.primary": {
11
+ "type": "VISUAL",
12
+ "shape": [
13
+ 3,
14
+ 256,
15
+ 256
16
+ ]
17
+ },
18
+ "observation.images.tactile": {
19
+ "type": "VISUAL",
20
+ "shape": [
21
+ 3,
22
+ 256,
23
+ 256
24
+ ]
25
+ },
26
+ "observation.images.wrist": {
27
+ "type": "VISUAL",
28
+ "shape": [
29
+ 3,
30
+ 256,
31
+ 256
32
+ ]
33
+ }
34
+ },
35
+ "output_features": {
36
+ "action": {
37
+ "type": "ACTION",
38
+ "shape": [
39
+ 7
40
+ ]
41
+ }
42
+ },
43
+ "device": "cuda",
44
+ "use_amp": false,
45
+ "push_to_hub": false,
46
+ "repo_id": "continuallearning/smolvla_image_only_lsy_plate",
47
+ "private": null,
48
+ "tags": null,
49
+ "license": null,
50
+ "chunk_size": 50,
51
+ "n_action_steps": 50,
52
+ "max_state_dim": 32,
53
+ "max_action_dim": 32,
54
+ "resize_imgs_with_padding": [
55
+ 512,
56
+ 512
57
+ ],
58
+ "empty_cameras": 0,
59
+ "adapt_to_pi_aloha": false,
60
+ "use_delta_joint_actions_aloha": false,
61
+ "tokenizer_max_length": 48,
62
+ "num_steps": 10,
63
+ "use_cache": true,
64
+ "freeze_vision_encoder": true,
65
+ "train_expert_only": true,
66
+ "train_state_proj": false,
67
+ "optimizer_lr": 0.0001,
68
+ "optimizer_betas": [
69
+ 0.9,
70
+ 0.95
71
+ ],
72
+ "optimizer_eps": 1e-08,
73
+ "optimizer_weight_decay": 1e-10,
74
+ "optimizer_grad_clip_norm": 10.0,
75
+ "scheduler_warmup_steps": 100,
76
+ "scheduler_decay_steps": 60000,
77
+ "scheduler_decay_lr": 2.5e-06,
78
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
79
+ "load_vlm_weights": true,
80
+ "add_image_special_tokens": false,
81
+ "attention_mode": "cross_attn",
82
+ "prefix_length": -1,
83
+ "pad_language_to": "longest",
84
+ "num_expert_layers": -1,
85
+ "num_vlm_layers": 16,
86
+ "self_attn_every_n_layers": 2,
87
+ "expert_width_multiplier": 0.75,
88
+ "min_period": 0.004,
89
+ "max_period": 4.0
90
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9297a7960f9a7c1ef886704cf9df93a24527d886b0438cef395ce8ff8dad45c0
3
+ size 906713040
train_config.json ADDED
@@ -0,0 +1,204 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "LSY-lab/stack_v2",
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_image_only",
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.primary": {
78
+ "type": "VISUAL",
79
+ "shape": [
80
+ 3,
81
+ 256,
82
+ 256
83
+ ]
84
+ },
85
+ "observation.images.tactile": {
86
+ "type": "VISUAL",
87
+ "shape": [
88
+ 3,
89
+ 256,
90
+ 256
91
+ ]
92
+ },
93
+ "observation.images.wrist": {
94
+ "type": "VISUAL",
95
+ "shape": [
96
+ 3,
97
+ 256,
98
+ 256
99
+ ]
100
+ }
101
+ },
102
+ "output_features": {
103
+ "action": {
104
+ "type": "ACTION",
105
+ "shape": [
106
+ 7
107
+ ]
108
+ }
109
+ },
110
+ "device": "cuda",
111
+ "use_amp": false,
112
+ "push_to_hub": false,
113
+ "repo_id": "continuallearning/smolvla_image_only_lsy_plate",
114
+ "private": null,
115
+ "tags": null,
116
+ "license": null,
117
+ "chunk_size": 50,
118
+ "n_action_steps": 50,
119
+ "max_state_dim": 32,
120
+ "max_action_dim": 32,
121
+ "resize_imgs_with_padding": [
122
+ 512,
123
+ 512
124
+ ],
125
+ "empty_cameras": 0,
126
+ "adapt_to_pi_aloha": false,
127
+ "use_delta_joint_actions_aloha": false,
128
+ "tokenizer_max_length": 48,
129
+ "num_steps": 10,
130
+ "use_cache": true,
131
+ "freeze_vision_encoder": true,
132
+ "train_expert_only": true,
133
+ "train_state_proj": false,
134
+ "optimizer_lr": 0.0001,
135
+ "optimizer_betas": [
136
+ 0.9,
137
+ 0.95
138
+ ],
139
+ "optimizer_eps": 1e-08,
140
+ "optimizer_weight_decay": 1e-10,
141
+ "optimizer_grad_clip_norm": 10.0,
142
+ "scheduler_warmup_steps": 100,
143
+ "scheduler_decay_steps": 60000,
144
+ "scheduler_decay_lr": 2.5e-06,
145
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
146
+ "load_vlm_weights": true,
147
+ "add_image_special_tokens": false,
148
+ "attention_mode": "cross_attn",
149
+ "prefix_length": -1,
150
+ "pad_language_to": "longest",
151
+ "num_expert_layers": -1,
152
+ "num_vlm_layers": 16,
153
+ "self_attn_every_n_layers": 2,
154
+ "expert_width_multiplier": 0.75,
155
+ "min_period": 0.004,
156
+ "max_period": 4.0
157
+ },
158
+ "output_dir": "outputs/train/2025-12-13/05-41-41_test_smolvla_image_only_lsy_drawer",
159
+ "job_name": "test_smolvla_image_only_lsy_drawer",
160
+ "resume": false,
161
+ "seed": 1000,
162
+ "num_workers": 1,
163
+ "batch_size": 1,
164
+ "steps": 10,
165
+ "eval_freq": 0,
166
+ "log_freq": 1,
167
+ "save_checkpoint": true,
168
+ "save_freq": 10,
169
+ "use_policy_training_preset": true,
170
+ "optimizer": {
171
+ "type": "adamw",
172
+ "lr": 0.0001,
173
+ "weight_decay": 1e-10,
174
+ "grad_clip_norm": 10.0,
175
+ "betas": [
176
+ 0.9,
177
+ 0.95
178
+ ],
179
+ "eps": 1e-08
180
+ },
181
+ "scheduler": {
182
+ "type": "cosine_decay_with_warmup",
183
+ "num_warmup_steps": 100,
184
+ "num_decay_steps": 60000,
185
+ "peak_lr": 0.0001,
186
+ "decay_lr": 2.5e-06
187
+ },
188
+ "eval": {
189
+ "n_episodes": 50,
190
+ "batch_size": 50,
191
+ "use_async_envs": false,
192
+ "max_episodes_rendered": 4
193
+ },
194
+ "wandb": {
195
+ "enable": true,
196
+ "disable_artifact": true,
197
+ "project": "debug",
198
+ "entity": "470620104-technical-university-of-munich",
199
+ "notes": null,
200
+ "run_id": "lwnp9cy5",
201
+ "mode": null
202
+ },
203
+ "eval_with_dataset": null
204
+ }